Up next in 10
The Cloud Show with Magnus Mårtensson ft. Anurag Karuparti & Paul S - Ep: 37
0 views
Aug 6, 2025
Join this live session with Magnus Mårtensson ft. Anurag Karupar & Paul S for the next episode of The Cloud Show with Magnus Mårtensson on July 24, at 01:05 PM (EST). The show is about cloud leadership and all the important questions relating to cloud projects. Certainly, many matters when a company is going to and wants to be successful in the cloud, are about technology. However, there are many additional matters, adjacent to technology, that we also need to tend to regarding business strategy, human resources, organizational change, planning for a technical cloud approach, and many more questions. These conversations are critical for a healthy cloud and for a swift and accurate cloud approach. GUEST SPEAKERS Anurag Karupar works in Microsoft's Data and AI group as a Senior AI Cloud Solution Architect where he advises Fortune 500 companies to design and build modern AI solutions using Azure AI products and services. Prior to joining Microsoft, he gained valuable experience at top consulting firms like PwC and EY, absorbing knowledge at every turn. His first book "Generative AI for Cloud Solutions" talks about his experiences and lessons learned while productionizing modern AI solutions for customers during the initial Gen AI boom. In 2024, Anurag was chosen to join the PaceSetters Leadership Program at Microsoft, a prestigious technical community dedicated to individuals committed to making a tangible impact through technology leadership. Paul S an individual with a passion for people and technology - a desire to make an impact with a commitment to challenging and rewarding work. He enjoys using my broad and deep skills to provide technical solutions to help clients, various industries, partners, and organizations succeed! 📺 CSharp TV - Dev Streaming Destination http://csharp.tv 🌎 C# Corner - Community of Software and Data Developers https://www.c-sharpcorner.com #CSharpTV #csharpcorner #TheCloudShow #CSharp #Interview
View Video Transcript
0:01
hello everyone and welcome back to
0:03
hello everyone and welcome back to
0:03
hello everyone and welcome back to another episode of the cloud show today
0:06
another episode of the cloud show today
0:06
another episode of the cloud show today I'm going to talk to Two Gentlemen who
0:08
I'm going to talk to Two Gentlemen who
0:08
I'm going to talk to Two Gentlemen who are actually I think it's maybe the
0:10
are actually I think it's maybe the
0:10
are actually I think it's maybe the first time I have two people on my show
0:12
first time I have two people on my show
0:12
first time I have two people on my show it's cool so we're GNA talk about
0:14
it's cool so we're GNA talk about
0:14
it's cool so we're GNA talk about something that I believe this industry
0:16
something that I believe this industry
0:16
something that I believe this industry right now can't get enough of talking
0:18
right now can't get enough of talking
0:18
right now can't get enough of talking about we are going to talk about
0:19
about we are going to talk about
0:19
about we are going to talk about generative Ai and two experts are with
0:23
generative Ai and two experts are with
0:23
generative Ai and two experts are with me on the cloud show today who have
0:25
me on the cloud show today who have
0:25
me on the cloud show today who have written a book on the topic and who
0:27
written a book on the topic and who
0:27
written a book on the topic and who works with this uh extensively and and
0:30
works with this uh extensively and and
0:30
works with this uh extensively and and and in detail so the experts the stars
0:32
and in detail so the experts the stars
0:32
and in detail so the experts the stars of the show tonight is Paul and
0:36
of the show tonight is Paul and
0:36
of the show tonight is Paul and [Music]
0:45
Anu hello gentlemen hello how are you
0:48
Anu hello gentlemen hello how are you
0:48
Anu hello gentlemen hello how are you hello hi Magnus hi everyone it's good
0:51
hello hi Magnus hi everyone it's good
0:51
hello hi Magnus hi everyone it's good it's good to have you on the show
0:52
it's good to have you on the show
0:52
it's good to have you on the show Welcome to the cloud show excited to be
0:55
Welcome to the cloud show excited to be
0:55
Welcome to the cloud show excited to be here thank you for having us brilant
0:57
here thank you for having us brilant
0:57
here thank you for having us brilant thank you yeah let's let's let's into
1:00
thank you yeah let's let's let's into
1:00
thank you yeah let's let's let's into into it uh straight away and and um int
1:04
into it uh straight away and and um int
1:04
into it uh straight away and and um int through you guys who are you and what
1:05
through you guys who are you and what
1:05
through you guys who are you and what what do you
1:06
what do you
1:06
what do you do sure thank you Magnus uh I I could
1:09
do sure thank you Magnus uh I I could
1:10
do sure thank you Magnus uh I I could start my name is Paul sing I'm a
1:11
start my name is Paul sing I'm a
1:11
start my name is Paul sing I'm a currently a principal Cloud solution
1:13
currently a principal Cloud solution
1:13
currently a principal Cloud solution architect with Microsoft covering data
1:16
architect with Microsoft covering data
1:16
architect with Microsoft covering data and AI this is a technical role so I
1:19
and AI this is a technical role so I
1:19
and AI this is a technical role so I provide guidance to our customers and
1:21
provide guidance to our customers and
1:21
provide guidance to our customers and organizations who are adopting the
1:23
organizations who are adopting the
1:23
organizations who are adopting the Microsoft Azure Cloud technology spe
1:26
Microsoft Azure Cloud technology spe
1:26
Microsoft Azure Cloud technology spe specifically around the data and AI
1:27
specifically around the data and AI
1:27
specifically around the data and AI domain um I focus on Professional
1:30
domain um I focus on Professional
1:30
domain um I focus on Professional Services companies primarily at this
1:32
Services companies primarily at this
1:32
Services companies primarily at this time so large Fortune 50 companies yeah
1:35
time so large Fortune 50 companies yeah
1:35
time so large Fortune 50 companies yeah yep y That's yeah I'm I'm I'm based in
1:38
yep y That's yeah I'm I'm I'm based in
1:38
yep y That's yeah I'm I'm I'm based in Northern California so you know lots of
1:40
Northern California so you know lots of
1:40
Northern California so you know lots of sunshine here sometimes too much and
1:43
sunshine here sometimes too much and
1:43
sunshine here sometimes too much and nice I'm currently working on also a
1:46
nice I'm currently working on also a
1:46
nice I'm currently working on also a side project for my personal time I'm
1:47
side project for my personal time I'm
1:48
side project for my personal time I'm volunteering for a nonprofit
1:49
volunteering for a nonprofit
1:49
volunteering for a nonprofit organization that's focused on stem uh
1:53
organization that's focused on stem uh
1:53
organization that's focused on stem uh growth they're called human bulb and
1:55
growth they're called human bulb and
1:55
growth they're called human bulb and what they do is they help
1:56
what they do is they help
1:56
what they do is they help underprivileged students and also
1:59
underprivileged students and also
1:59
underprivileged students and also business leaders on better understanding
2:01
business leaders on better understanding
2:01
business leaders on better understanding Technologies and how to adopt you know
2:04
Technologies and how to adopt you know
2:04
Technologies and how to adopt you know new technologies like gen for example so
2:06
new technologies like gen for example so
2:06
new technologies like gen for example so fantastic that's lovely okay cool
2:08
fantastic that's lovely okay cool
2:08
fantastic that's lovely okay cool brilliant and Anu what's up with you yes
2:12
brilliant and Anu what's up with you yes
2:12
brilliant and Anu what's up with you yes similar to Paul I'm I'm a senior Cloud
2:14
similar to Paul I'm I'm a senior Cloud
2:14
similar to Paul I'm I'm a senior Cloud solution architect specializing in data
2:15
solution architect specializing in data
2:15
solution architect specializing in data and AI based out of New York and it's
2:18
and AI based out of New York and it's
2:18
and AI based out of New York and it's quite hot here
2:20
quite hot here
2:20
quite hot here today and uh yeah so as a cloud
2:23
today and uh yeah so as a cloud
2:23
today and uh yeah so as a cloud architect uh so I'm helping the the SE
2:26
architect uh so I'm helping the the SE
2:26
architect uh so I'm helping the the SE Suite Executives on the data and
2:28
Suite Executives on the data and
2:28
Suite Executives on the data and strategy one day the next day it would
2:30
strategy one day the next day it would
2:30
strategy one day the next day it would be going more lower level with the
2:32
be going more lower level with the
2:32
be going more lower level with the engineering teams right uh helping them
2:34
engineering teams right uh helping them
2:34
engineering teams right uh helping them unblock on technical stuff and then also
2:37
unblock on technical stuff and then also
2:37
unblock on technical stuff and then also upskilling them on Azure AI products and
2:39
upskilling them on Azure AI products and
2:39
upskilling them on Azure AI products and services my goal is to ensure that they
2:42
services my goal is to ensure that they
2:42
services my goal is to ensure that they are successful on Azure products and
2:43
are successful on Azure products and
2:43
are successful on Azure products and services data and products and services
2:46
services data and products and services
2:46
services data and products and services so that's uh about my role apart from
2:49
so that's uh about my role apart from
2:49
so that's uh about my role apart from work I like running marathons uh and
2:52
work I like running marathons uh and
2:52
work I like running marathons uh and also I Mentor students in Africa like
2:56
also I Mentor students in Africa like
2:56
also I Mentor students in Africa like through the global mentorship initiative
2:58
through the global mentorship initiative
2:58
through the global mentorship initiative uh like college students so so that's a
3:00
uh like college students so so that's a
3:00
uh like college students so so that's a little about me that's lovely lovely
3:03
little about me that's lovely lovely
3:03
little about me that's lovely lovely absolutely lovely okay brilliant well
3:05
absolutely lovely okay brilliant well
3:05
absolutely lovely okay brilliant well it's it's it's wonderful to have you
3:06
it's it's it's wonderful to have you
3:06
it's it's it's wonderful to have you guys on this show thank you for being
3:08
guys on this show thank you for being
3:08
guys on this show thank you for being part of the cloud show and and of course
3:10
part of the cloud show and and of course
3:11
part of the cloud show and and of course we're going to talk uh about your book
3:12
we're going to talk uh about your book
3:12
we're going to talk uh about your book but I'd like to talk about what went
3:14
but I'd like to talk about what went
3:14
but I'd like to talk about what went into your book uh we're want to talk
3:16
into your book uh we're want to talk
3:16
into your book uh we're want to talk about um you know generative AI
3:18
about um you know generative AI
3:18
about um you know generative AI Solutions architecture that sort of
3:20
Solutions architecture that sort of
3:20
Solutions architecture that sort of stuff um you know what are the like the
3:23
stuff um you know what are the like the
3:23
stuff um you know what are the like the if you will they like the key components
3:25
if you will they like the key components
3:25
if you will they like the key components of this hey we have a sign let's let's
3:28
of this hey we have a sign let's let's
3:28
of this hey we have a sign let's let's show the book Let's Do It Let's do it
3:29
show the book Let's Do It Let's do it
3:30
show the book Let's Do It Let's do it straight away this is the book you see
3:31
straight away this is the book you see
3:31
straight away this is the book you see the names all right I see the names and
3:36
the names all right I see the names and
3:36
the names all right I see the names and oh he made it larger good all right so
3:39
oh he made it larger good all right so
3:39
oh he made it larger good all right so there there's the book generative AI for
3:40
there there's the book generative AI for
3:40
there there's the book generative AI for cloud Solutions so let's talk about that
3:44
cloud Solutions so let's talk about that
3:44
cloud Solutions so let's talk about that like what is this uh when you're
3:46
like what is this uh when you're
3:46
like what is this uh when you're building generative AI Solutions and
3:48
building generative AI Solutions and
3:48
building generative AI Solutions and you're architecting this what goes into
3:50
you're architecting this what goes into
3:50
you're architecting this what goes into something like
3:52
something like
3:52
something like that sure any do you want to
3:55
that sure any do you want to
3:55
that sure any do you want to start sure uh so yeah so this is this
3:58
start sure uh so yeah so this is this
3:58
start sure uh so yeah so this is this book is basically it incorporates our
4:00
book is basically it incorporates our
4:00
book is basically it incorporates our experiences from the field but when we
4:02
experiences from the field but when we
4:02
experiences from the field but when we were directly working with the customers
4:04
were directly working with the customers
4:04
were directly working with the customers right so last one and a half years have
4:06
right so last one and a half years have
4:06
right so last one and a half years have been like extremely rewarding for us in
4:08
been like extremely rewarding for us in
4:08
been like extremely rewarding for us in terms of learning like moving so many
4:11
terms of learning like moving so many
4:11
terms of learning like moving so many generative solutions to production like
4:13
generative solutions to production like
4:14
generative solutions to production like I I I have like 23 customers across
4:16
I I I have like 23 customers across
4:16
I I I have like 23 customers across media entertainment and real estate uh
4:18
media entertainment and real estate uh
4:19
media entertainment and real estate uh helped around 50 plus use cases we
4:21
helped around 50 plus use cases we
4:21
helped around 50 plus use cases we guided them to production and so this is
4:24
guided them to production and so this is
4:24
guided them to production and so this is this book consists of all our
4:27
this book consists of all our
4:27
this book consists of all our experiences Lessons Learned and how do
4:29
experiences Lessons Learned and how do
4:29
experiences Lessons Learned and how do you build that robust architecture right
4:32
you build that robust architecture right
4:32
you build that robust architecture right gen architecture we also spoken from the
4:36
gen architecture we also spoken from the
4:36
gen architecture we also spoken from the core principle standpoint the first
4:37
core principle standpoint the first
4:38
core principle standpoint the first principles like rag fine tuning when do
4:40
principles like rag fine tuning when do
4:40
principles like rag fine tuning when do you use it right you don't need it all
4:42
you use it right you don't need it all
4:42
you use it right you don't need it all the time when do you use it one over the
4:44
the time when do you use it one over the
4:44
the time when do you use it one over the other different typ of model we spoke
4:46
other different typ of model we spoke
4:46
other different typ of model we spoke about multimodal uh yeah and how to
4:49
about multimodal uh yeah and how to
4:49
about multimodal uh yeah and how to operationalize it and also yeah yeah
4:53
operationalize it and also yeah yeah
4:53
operationalize it and also yeah yeah Paul you want to add yeah yeah sure of
4:56
Paul you want to add yeah yeah sure of
4:56
Paul you want to add yeah yeah sure of course uh of course with any technical
4:58
course uh of course with any technical
4:58
course uh of course with any technical Endeavor for an organization there's the
5:00
Endeavor for an organization there's the
5:00
Endeavor for an organization there's the business strategy that's involved so we
5:02
business strategy that's involved so we
5:02
business strategy that's involved so we want to you know talk about some of the
5:05
want to you know talk about some of the
5:05
want to you know talk about some of the areas where there's organizational
5:07
areas where there's organizational
5:07
areas where there's organizational changes that occur when adopting new
5:09
changes that occur when adopting new
5:09
changes that occur when adopting new technologies such as gen AI yeah and
5:12
technologies such as gen AI yeah and
5:12
technologies such as gen AI yeah and this is moving fast now right I mean
5:14
this is moving fast now right I mean
5:14
this is moving fast now right I mean it's it's generative AI it's moving
5:17
it's it's generative AI it's moving
5:17
it's it's generative AI it's moving speed of light faster even it's so fast
5:20
speed of light faster even it's so fast
5:20
speed of light faster even it's so fast and it's growing so much and and and
5:22
and it's growing so much and and and
5:22
and it's growing so much and and and everybody I guess wants to get in on
5:25
everybody I guess wants to get in on
5:25
everybody I guess wants to get in on that like if you had a product before
5:27
that like if you had a product before
5:27
that like if you had a product before and if it doesn't have the product but
5:30
and if it doesn't have the product but
5:30
and if it doesn't have the product but now with AI you're not part you're not
5:32
now with AI you're not part you're not
5:32
now with AI you're not part you're not on the you're not on the train right
5:34
on the you're not on the train right
5:34
on the you're not on the train right it's like you're at the
5:36
it's like you're at the
5:36
it's like you're at the station yeah 100% I mean we see every
5:39
station yeah 100% I mean we see every
5:39
station yeah 100% I mean we see every organization trying to at least
5:40
organization trying to at least
5:40
organization trying to at least understand and learn generative Ai and
5:43
understand and learn generative Ai and
5:43
understand and learn generative Ai and trying to adopt it into the business
5:44
trying to adopt it into the business
5:44
trying to adopt it into the business processes I mean there's huge rewards
5:46
processes I mean there's huge rewards
5:46
processes I mean there's huge rewards there's a lot of companies that have
5:48
there's a lot of companies that have
5:48
there's a lot of companies that have done some the return on investment as
5:51
done some the return on investment as
5:51
done some the return on investment as well to
5:52
well to
5:52
well to yeah can make their users their internal
5:57
yeah can make their users their internal
5:57
yeah can make their users their internal more productive yeah so I guess that's
5:59
more productive yeah so I guess that's
5:59
more productive yeah so I guess that's that's part of it I
6:01
that's part of it I
6:01
that's part of it I mean using a service that somebody else
6:04
mean using a service that somebody else
6:04
mean using a service that somebody else created that is um that has added AI
6:08
created that is um that has added AI
6:08
created that is um that has added AI capabilities like you know Office 365
6:10
capabilities like you know Office 365
6:10
capabilities like you know Office 365 you could use co-pilots and things right
6:13
you could use co-pilots and things right
6:13
you could use co-pilots and things right um that's one thing but taking something
6:17
um that's one thing but taking something
6:17
um that's one thing but taking something AI into your own like organization and
6:19
AI into your own like organization and
6:19
AI into your own like organization and building something for you based on
6:22
building something for you based on
6:22
building something for you based on whatever business you're doing that's
6:24
whatever business you're doing that's
6:24
whatever business you're doing that's something else entirely um and and and
6:27
something else entirely um and and and
6:27
something else entirely um and and and how do you go about as a business to
6:31
how do you go about as a business to
6:31
how do you go about as a business to understand uh like we should also have
6:33
understand uh like we should also have
6:33
understand uh like we should also have some AI I guess because everybody says
6:35
some AI I guess because everybody says
6:35
some AI I guess because everybody says that everybody has to have that and what
6:36
that everybody has to have that and what
6:36
that everybody has to have that and what should we do
6:38
should we do
6:38
should we do now what's the what's the cat how do you
6:41
now what's the what's the cat how do you
6:41
now what's the what's the cat how do you get going with
6:43
get going with
6:43
get going with this so so what I would I mean other
6:46
this so so what I would I mean other
6:46
this so so what I would I mean other than buying the book and and reading
6:48
than buying the book and and reading
6:48
than buying the book and and reading through because we did sign we our
6:51
through because we did sign we our
6:51
through because we did sign we our approach was to you know that new and
6:55
approach was to you know that new and
6:55
approach was to you know that new and understanding but primarily it's like if
6:56
understanding but primarily it's like if
6:56
understanding but primarily it's like if you don't understand the technology you
6:58
you don't understand the technology you
6:58
you don't understand the technology you don't have to be an expert at it however
7:00
don't have to be an expert at it however
7:00
don't have to be an expert at it however I would recommend getting that basic
7:02
I would recommend getting that basic
7:02
I would recommend getting that basic understanding YouTube videos or other
7:04
understanding YouTube videos or other
7:04
understanding YouTube videos or other platforms there's plenty of free
7:06
platforms there's plenty of free
7:06
platforms there's plenty of free documents out there videos that are
7:08
documents out there videos that are
7:08
documents out there videos that are available um watching Cloud shows you
7:10
available um watching Cloud shows you
7:10
available um watching Cloud shows you know of course um definely yeah so
7:13
know of course um definely yeah so
7:13
know of course um definely yeah so that's understanding the the underlying
7:17
that's understanding the the underlying
7:17
that's understanding the the underlying principles is so so is there sort of a
7:19
principles is so so is there sort of a
7:19
principles is so so is there sort of a structured approach for a a a company to
7:22
structured approach for a a a company to
7:22
structured approach for a a a company to like okay we we want to figure out AI
7:25
like okay we we want to figure out AI
7:25
like okay we we want to figure out AI for us what it means for us for our
7:27
for us what it means for us for our
7:27
for us what it means for us for our business like how do you approach
7:29
business like how do you approach
7:29
business like how do you approach approach that is there like a breakdown
7:31
approach that is there like a breakdown
7:31
approach that is there like a breakdown thing you can do or
7:34
thing you can do or
7:34
thing you can do or yeah so yeah I think the first thing is
7:37
yeah so yeah I think the first thing is
7:37
yeah so yeah I think the first thing is to go do like a comprehensive ideation
7:39
to go do like a comprehensive ideation
7:39
to go do like a comprehensive ideation phase right and what we observed was
7:42
phase right and what we observed was
7:42
phase right and what we observed was hackathons were very successful they
7:44
hackathons were very successful they
7:44
hackathons were very successful they were like breeding grounds for Creative
7:46
were like breeding grounds for Creative
7:46
were like breeding grounds for Creative Solutions innovative solutions it got a
7:48
Solutions innovative solutions it got a
7:48
Solutions innovative solutions it got a lot of innovation out of the employees
7:51
lot of innovation out of the employees
7:51
lot of innovation out of the employees but you will see so many use cases come
7:53
but you will see so many use cases come
7:53
but you will see so many use cases come out of these hackathons you don't want
7:55
out of these hackathons you don't want
7:55
out of these hackathons you don't want to do everything you just want to narrow
7:57
to do everything you just want to narrow
7:57
to do everything you just want to narrow down on few of them that add value to
8:00
down on few of them that add value to
8:00
down on few of them that add value to your to your industry to your business
8:03
your to your industry to your business
8:03
your to your industry to your business right that may be productivity boost uh
8:06
right that may be productivity boost uh
8:06
right that may be productivity boost uh so one of the things we saw was the
8:08
so one of the things we saw was the
8:08
so one of the things we saw was the internal co-pilots like using Azure
8:09
internal co-pilots like using Azure
8:10
internal co-pilots like using Azure opener on your data that was a big hit
8:12
opener on your data that was a big hit
8:12
opener on your data that was a big hit and those Solutions picked up because it
8:15
and those Solutions picked up because it
8:15
and those Solutions picked up because it was like quick time to Val to Quick time
8:17
was like quick time to Val to Quick time
8:17
was like quick time to Val to Quick time to Market and like it was quick wins and
8:20
to Market and like it was quick wins and
8:20
to Market and like it was quick wins and it gave immediate immediate value and
8:22
it gave immediate immediate value and
8:22
it gave immediate immediate value and and we we saw that a lot mhm so
8:25
and we we saw that a lot mhm so
8:25
and we we saw that a lot mhm so essentially hack away like gather people
8:29
essentially hack away like gather people
8:29
essentially hack away like gather people hack away take some of your data and and
8:31
hack away take some of your data and and
8:31
hack away take some of your data and and try to like do something not production
8:35
try to like do something not production
8:35
try to like do something not production product yet but figure out what to do
8:38
product yet but figure out what to do
8:38
product yet but figure out what to do corre get get comfortable with the tools
8:40
corre get get comfortable with the tools
8:40
corre get get comfortable with the tools and the development process and if
8:42
and the development process and if
8:42
and the development process and if there's any gaps in the organization
8:44
there's any gaps in the organization
8:44
there's any gaps in the organization such as knowledge training then you
8:45
such as knowledge training then you
8:46
such as knowledge training then you identify those during the hackathon okay
8:48
identify those during the hackathon okay
8:48
identify those during the hackathon okay okay and so what when when companies are
8:51
okay and so what when when companies are
8:51
okay and so what when when companies are doing like you like you've mentioned
8:53
doing like you like you've mentioned
8:53
doing like you like you've mentioned here you have been doing this with a lot
8:55
here you have been doing this with a lot
8:55
here you have been doing this with a lot of companies and and a lot of projects
8:58
of companies and and a lot of projects
8:58
of companies and and a lot of projects um what are some typical
9:00
um what are some typical
9:00
um what are some typical challenges pitfalls problems that that
9:03
challenges pitfalls problems that that
9:03
challenges pitfalls problems that that that that your customers run into what
9:05
that that your customers run into what
9:05
that that your customers run into what what is hard here except learning the
9:08
what is hard here except learning the
9:08
what is hard here except learning the technology I guess but that's you know
9:10
technology I guess but that's you know
9:10
technology I guess but that's you know that's learning yeah well that that's
9:13
that's learning yeah well that that's
9:13
that's learning yeah well that that's exactly it is I mean so some of is human
9:15
exactly it is I mean so some of is human
9:15
exactly it is I mean so some of is human resource component as well make sure you
9:17
resource component as well make sure you
9:17
resource component as well make sure you have the right staff some have have to
9:19
have the right staff some have have to
9:19
have the right staff some have have to have developer background because you
9:21
have developer background because you
9:21
have developer background because you are developing against your own data
9:23
are developing against your own data
9:23
are developing against your own data although you leverag AI in the cloud but
9:26
although you leverag AI in the cloud but
9:26
although you leverag AI in the cloud but there's still development against your
9:28
there's still development against your
9:28
there's still development against your own data um I would also some of the
9:31
own data um I would also some of the
9:31
own data um I would also some of the challenge is like things understanding
9:34
challenge is like things understanding
9:34
challenge is like things understanding the Model Behavior they're called large
9:36
the Model Behavior they're called large
9:36
the Model Behavior they're called large language models and each model has its
9:38
language models and each model has its
9:38
language models and each model has its own limitations on how much data it can
9:41
own limitations on how much data it can
9:41
own limitations on how much data it can process so if you're an organization and
9:43
process so if you're an organization and
9:43
process so if you're an organization and you want to throw a large number of
9:44
you want to throw a large number of
9:44
you want to throw a large number of queries questions we call it prompts
9:46
queries questions we call it prompts
9:46
queries questions we call it prompts against a model just understand that
9:49
against a model just understand that
9:49
against a model just understand that there are limitations on any Cloud
9:52
there are limitations on any Cloud
9:52
there are limitations on any Cloud resource so people are not understanding
9:55
resource so people are not understanding
9:55
resource so people are not understanding how to calibrate yet not initially no
9:58
how to calibrate yet not initially no
9:58
how to calibrate yet not initially no and this it's a iterative process so
10:00
and this it's a iterative process so
10:00
and this it's a iterative process so there's a lot of testing that goes on
10:01
there's a lot of testing that goes on
10:01
there's a lot of testing that goes on against their own data and against
10:03
against their own data and against
10:03
against their own data and against different types of models and newer
10:05
different types of models and newer
10:05
different types of models and newer models come out every day that are more
10:07
models come out every day that are more
10:07
models come out every day that are more efficient lower cost as well so there's
10:10
efficient lower cost as well so there's
10:10
efficient lower cost as well so there's you know some
10:12
you know some
10:12
you know some it all right so learn about these things
10:14
it all right so learn about these things
10:14
it all right so learn about these things and start um testing and and and you
10:17
and start um testing and and and you
10:17
and start um testing and and and you know iteratively try out new things new
10:19
know iteratively try out new things new
10:19
know iteratively try out new things new models different approaches correct yeah
10:22
models different approaches correct yeah
10:22
models different approaches correct yeah yeah and one more thing is like well so
10:24
yeah and one more thing is like well so
10:24
yeah and one more thing is like well so we saw a lot of rag based Solutions
10:26
we saw a lot of rag based Solutions
10:26
we saw a lot of rag based Solutions going into production and now what we're
10:28
going into production and now what we're
10:28
going into production and now what we're seeing is the customer expectations are
10:30
seeing is the customer expectations are
10:30
seeing is the customer expectations are rising on the quality of the outputs of
10:33
rising on the quality of the outputs of
10:33
rising on the quality of the outputs of these rag Solutions so what I would
10:35
these rag Solutions so what I would
10:36
these rag Solutions so what I would suggest is like having like an
10:37
suggest is like having like an
10:37
suggest is like having like an evaluation strategy early on that can
10:40
evaluation strategy early on that can
10:40
evaluation strategy early on that can really help you see surprises uh help
10:43
really help you see surprises uh help
10:43
really help you see surprises uh help you stop seeing surprises later on right
10:46
you stop seeing surprises later on right
10:46
you stop seeing surprises later on right so like doing uh initial assessment with
10:48
so like doing uh initial assessment with
10:48
so like doing uh initial assessment with different models and like having a
10:50
different models and like having a
10:50
different models and like having a qualitative quantitative metrics to
10:53
qualitative quantitative metrics to
10:53
qualitative quantitative metrics to assess the quality of the outputs that
10:55
assess the quality of the outputs that
10:55
assess the quality of the outputs that really helps early
10:56
really helps early
10:56
really helps early on okay okay all right that makes sense
10:59
on okay okay all right that makes sense
10:59
on okay okay all right that makes sense that makes sense it's it's something
11:01
that makes sense it's it's something
11:01
that makes sense it's it's something that you have to get into and and really
11:02
that you have to get into and and really
11:02
that you have to get into and and really start doing I guess to to really fully
11:05
start doing I guess to to really fully
11:05
start doing I guess to to really fully understand um this process but but
11:07
understand um this process but but
11:07
understand um this process but but essentially you're saying start doing
11:10
essentially you're saying start doing
11:10
essentially you're saying start doing something and and and then try to qu you
11:13
something and and and then try to qu you
11:13
something and and and then try to qu you know qualify it get better and and more
11:16
know qualify it get better and and more
11:16
know qualify it get better and and more focused on what you should be doing but
11:18
focused on what you should be doing but
11:18
focused on what you should be doing but you don't know from the beginning so at
11:19
you don't know from the beginning so at
11:19
you don't know from the beginning so at least do
11:20
least do
11:20
least do something let me add there's a lot of uh
11:23
something let me add there's a lot of uh
11:23
something let me add there's a lot of uh Cloud vendors that provide this
11:24
Cloud vendors that provide this
11:24
Cloud vendors that provide this playground environment where you can
11:26
playground environment where you can
11:26
playground environment where you can test against these models like Asher for
11:28
test against these models like Asher for
11:29
test against these models like Asher for example
11:29
example
11:29
example yeah definitely of course Azure Azure
11:32
yeah definitely of course Azure Azure
11:32
yeah definitely of course Azure Azure open AI you know there's the AI we're
11:34
open AI you know there's the AI we're
11:34
open AI you know there's the AI we're allowed to say
11:37
it okay brilliant so um since this space
11:40
it okay brilliant so um since this space
11:40
it okay brilliant so um since this space is kind of moving so fast because it
11:43
is kind of moving so fast because it
11:43
is kind of moving so fast because it really is there's new models coming out
11:45
really is there's new models coming out
11:45
really is there's new models coming out all the time it seems and and things are
11:47
all the time it seems and and things are
11:47
all the time it seems and and things are happening at at a at an not an alarming
11:49
happening at at a at an not an alarming
11:49
happening at at a at an not an alarming rate but a very very high rate um what
11:53
rate but a very very high rate um what
11:53
rate but a very very high rate um what are like the latest uh Trends what's
11:55
are like the latest uh Trends what's
11:55
are like the latest uh Trends what's going on right now in generative AI yeah
11:59
going on right now in generative AI yeah
11:59
going on right now in generative AI yeah K start and then I know an's got a few
12:00
K start and then I know an's got a few
12:00
K start and then I know an's got a few as well but uh so Anu mentioned rag
12:03
as well but uh so Anu mentioned rag
12:03
as well but uh so Anu mentioned rag which is retrieval augmented generation
12:05
which is retrieval augmented generation
12:05
which is retrieval augmented generation and this is using large langage models
12:07
and this is using large langage models
12:07
and this is using large langage models against your own data you know that's
12:09
against your own data you know that's
12:09
against your own data you know that's been around uh but to increase the speed
12:13
been around uh but to increase the speed
12:13
been around uh but to increase the speed and also approve in increase the
12:17
and also approve in increase the
12:17
and also approve in increase the response uh there's newer Technologies
12:19
response uh there's newer Technologies
12:19
response uh there's newer Technologies coming out such as integration with
12:21
coming out such as integration with
12:21
coming out such as integration with knowledge graphs so new and great such
12:25
knowledge graphs so new and great such
12:25
knowledge graphs so new and great such as just AI graphs a knowledge graph is
12:28
as just AI graphs a knowledge graph is
12:28
as just AI graphs a knowledge graph is it's what it is instead of using vectors
12:30
it's what it is instead of using vectors
12:30
it's what it is instead of using vectors which rag uses uh predominantly
12:32
which rag uses uh predominantly
12:32
which rag uses uh predominantly currently the it's an intersection
12:34
currently the it's an intersection
12:34
currently the it's an intersection between ji and Knowledge Graph so
12:37
between ji and Knowledge Graph so
12:37
between ji and Knowledge Graph so Knowledge Graph can gen and what and
12:40
Knowledge Graph can gen and what and
12:40
Knowledge Graph can gen and what and knowledge knowledge graphs yeah what
12:42
knowledge knowledge graphs yeah what
12:42
knowledge knowledge graphs yeah what these knowledge graphs do is they
12:44
these knowledge graphs do is they
12:44
these knowledge graphs do is they connect various data points and shows
12:46
connect various data points and shows
12:46
connect various data points and shows the relationship between those data
12:48
the relationship between those data
12:48
the relationship between those data within an
12:50
within an
12:50
within an organization that the outcome is more
12:52
organization that the outcome is more
12:52
organization that the outcome is more accurate results uh because if you do
12:55
accurate results uh because if you do
12:55
accurate results uh because if you do some prom sometimes you don't get the
12:56
some prom sometimes you don't get the
12:56
some prom sometimes you don't get the accurate results You're Expecting so
12:58
accurate results You're Expecting so
12:58
accurate results You're Expecting so we're seeing a lot
12:59
we're seeing a lot
12:59
we're seeing a lot a movement toward using more of a AI
13:02
a movement toward using more of a AI
13:03
a movement toward using more of a AI plus graph knowledge graph solutions oh
13:05
plus graph knowledge graph solutions oh
13:05
plus graph knowledge graph solutions oh interesting interesting so and and just
13:07
interesting interesting so and and just
13:07
interesting interesting so and and just so happens that Microsoft has one of the
13:10
so happens that Microsoft has one of the
13:10
so happens that Microsoft has one of the largest graph database uh products that
13:12
largest graph database uh products that
13:13
largest graph database uh products that exist uh or maybe the largest right
13:15
exist uh or maybe the largest right
13:15
exist uh or maybe the largest right Cosmos yeah oh yeah and as a matter of
13:18
Cosmos yeah oh yeah and as a matter of
13:18
Cosmos yeah oh yeah and as a matter of fact there's a a new service Cosmos AI
13:20
fact there's a a new service Cosmos AI
13:20
fact there's a a new service Cosmos AI graph that is based off of the knowledge
13:23
graph that is based off of the knowledge
13:23
graph that is based off of the knowledge graph plus AI that was not a leading
13:25
graph plus AI that was not a leading
13:25
graph plus AI that was not a leading question at all not not not even
13:29
thanks
13:31
thanks
13:31
thanks Magnus no but it's important right
13:33
Magnus no but it's important right
13:33
Magnus no but it's important right everything is being adapted to uh being
13:36
everything is being adapted to uh being
13:36
everything is being adapted to uh being able to handle these types of services
13:39
able to handle these types of services
13:39
able to handle these types of services and demands um so every Service uh is
13:43
and demands um so every Service uh is
13:43
and demands um so every Service uh is adapted to being able to to take care of
13:46
adapted to being able to to take care of
13:46
adapted to being able to to take care of this with with speed and accuracy and
13:48
this with with speed and accuracy and
13:48
this with with speed and accuracy and and you know good money right can't be
13:51
and you know good money right can't be
13:51
and you know good money right can't be too expensive as well and efficient
13:53
too expensive as well and efficient
13:53
too expensive as well and efficient correct yeah efficiency yes we have a
13:55
correct yeah efficiency yes we have a
13:55
correct yeah efficiency yes we have a few others on you want to mention yeah
13:58
few others on you want to mention yeah
13:58
few others on you want to mention yeah my my do one is autonomous agents we
14:00
my my do one is autonomous agents we
14:00
my my do one is autonomous agents we seeing like a lot of progress in this
14:02
seeing like a lot of progress in this
14:02
seeing like a lot of progress in this area so autonomous agents are basically
14:05
area so autonomous agents are basically
14:05
area so autonomous agents are basically these are intelligent AI systems that
14:07
these are intelligent AI systems that
14:07
these are intelligent AI systems that can act independently right with little
14:10
can act independently right with little
14:10
can act independently right with little or no human intervention they can uh do
14:13
or no human intervention they can uh do
14:13
or no human intervention they can uh do self feedback and improve automatically
14:15
self feedback and improve automatically
14:15
self feedback and improve automatically right so these I think will have massive
14:19
right so these I think will have massive
14:19
right so these I think will have massive productivity boost because of this and
14:21
productivity boost because of this and
14:21
productivity boost because of this and some of the use cases I see disrupting
14:24
some of the use cases I see disrupting
14:24
some of the use cases I see disrupting is like call centers you'll have like a
14:25
is like call centers you'll have like a
14:26
is like call centers you'll have like a team of autonomous agents that can
14:29
team of autonomous agents that can
14:29
team of autonomous agents that can augment human capabilities and like make
14:31
augment human capabilities and like make
14:31
augment human capabilities and like make them more productive so that's one I
14:34
them more productive so that's one I
14:34
them more productive so that's one I think it it could be helpful even like
14:35
think it it could be helpful even like
14:35
think it it could be helpful even like in education uh it could provide more
14:38
in education uh it could provide more
14:38
in education uh it could provide more personalized education to students so
14:40
personalized education to students so
14:41
personalized education to students so with like having a team of autonomous
14:42
with like having a team of autonomous
14:42
with like having a team of autonomous agents so I think that is going to be
14:45
agents so I think that is going to be
14:45
agents so I think that is going to be big and also like the market for
14:46
big and also like the market for
14:46
big and also like the market for autonomous agents is expected to grow to
14:48
autonomous agents is expected to grow to
14:48
autonomous agents is expected to grow to like close to $30 billion by 2028 in
14:51
like close to $30 billion by 2028 in
14:51
like close to $30 billion by 2028 in North America alone so that is one the
14:54
North America alone so that is one the
14:54
North America alone so that is one the second one is multimodal like we saw a
14:56
second one is multimodal like we saw a
14:56
second one is multimodal like we saw a lot of text based uh text based
15:00
lot of text based uh text based
15:00
lot of text based uh text based Solutions generative solution now it's
15:01
Solutions generative solution now it's
15:01
Solutions generative solution now it's becoming images audio and then soon
15:03
becoming images audio and then soon
15:03
becoming images audio and then soon there'll be video generation models also
15:05
there'll be video generation models also
15:05
there'll be video generation models also right and then small small language
15:07
right and then small small language
15:07
right and then small small language models is one more I I think that the
15:10
models is one more I I think that the
15:10
models is one more I I think that the trend we are seeing which is going to be
15:12
trend we are seeing which is going to be
15:12
trend we are seeing which is going to be big for Edge devices like smartphones
15:15
big for Edge devices like smartphones
15:15
big for Edge devices like smartphones and autonomous way oh yeah so having
15:17
and autonomous way oh yeah so having
15:17
and autonomous way oh yeah so having having a a small small large language
15:23
model yeah in the palm of your hand all
15:26
model yeah in the palm of your hand all
15:26
model yeah in the palm of your hand all right app Apple's working on that in
15:28
right app Apple's working on that in
15:28
right app Apple's working on that in their next
15:29
their next
15:29
their next IOS as well I think they are right huge
15:33
IOS as well I think they are right huge
15:33
IOS as well I think they are right huge it's funny how everyone everyone doesn't
15:35
it's funny how everyone everyone doesn't
15:35
it's funny how everyone everyone doesn't matter which company you are you you
15:37
matter which company you are you you
15:37
matter which company you are you you release something called whatever it was
15:39
release something called whatever it was
15:39
release something called whatever it was before plus AI in in some way shape or
15:42
before plus AI in in some way shape or
15:42
before plus AI in in some way shape or form
15:44
form
15:44
form right maybe known as the cloud show plus
15:46
right maybe known as the cloud show plus
15:46
right maybe known as the cloud show plus AI maybe who knows Magnus I I don't know
15:49
AI maybe who knows Magnus I I don't know
15:49
AI maybe who knows Magnus I I don't know like we couldn't possibly fit AI into no
15:52
like we couldn't possibly fit AI into no
15:52
like we couldn't possibly fit AI into no I'm
15:54
I'm
15:54
I'm kidding all right so
15:57
kidding all right so
15:57
kidding all right so um one one thing that that is um like so
16:01
um one one thing that that is um like so
16:01
um one one thing that that is um like so important as well in this space is
16:03
important as well in this space is
16:03
important as well in this space is around uh how do you how do you take
16:05
around uh how do you how do you take
16:05
around uh how do you how do you take care of your AI solution and and how do
16:07
care of your AI solution and and how do
16:07
care of your AI solution and and how do you ensure that you are uh behaving
16:11
you ensure that you are uh behaving
16:11
you ensure that you are uh behaving fairly I guess the word is ethically um
16:14
fairly I guess the word is ethically um
16:14
fairly I guess the word is ethically um this is of course a huge um concern um
16:17
this is of course a huge um concern um
16:17
this is of course a huge um concern um so what does it take for an organization
16:19
so what does it take for an organization
16:19
so what does it take for an organization to ensure that they are being ethical in
16:23
to ensure that they are being ethical in
16:23
to ensure that they are being ethical in their AI approach and what does that
16:24
their AI approach and what does that
16:24
their AI approach and what does that mean even yeah great question because
16:27
mean even yeah great question because
16:27
mean even yeah great question because you're right it is very critical having
16:29
you're right it is very critical having
16:29
you're right it is very critical having a responsible AI if you would um
16:31
a responsible AI if you would um
16:31
a responsible AI if you would um responsible AI components such as
16:33
responsible AI components such as
16:33
responsible AI components such as fairness reliability and safety privacy
16:37
fairness reliability and safety privacy
16:37
fairness reliability and safety privacy and security right they're all critical
16:39
and security right they're all critical
16:39
and security right they're all critical components in the overall gen strategy
16:41
components in the overall gen strategy
16:41
components in the overall gen strategy so when I talk to our customers we
16:43
so when I talk to our customers we
16:44
so when I talk to our customers we that's always a topic that we bring up
16:46
that's always a topic that we bring up
16:46
that's always a topic that we bring up right away so but I mean Cloud vendors
16:49
right away so but I mean Cloud vendors
16:49
right away so but I mean Cloud vendors like Microsoft have built- in support
16:51
like Microsoft have built- in support
16:51
like Microsoft have built- in support for many of these fun key functionality
16:53
for many of these fun key functionality
16:53
for many of these fun key functionality out of the box and it's very
16:55
out of the box and it's very
16:55
out of the box and it's very configurable as well so if you want to
16:56
configurable as well so if you want to
16:56
configurable as well so if you want to configure it for your organization can
16:59
configure it for your organization can
16:59
configure it for your organization can an actually expert in the raai space as
17:02
an actually expert in the raai space as
17:02
an actually expert in the raai space as well so that so I I I recently published
17:05
well so that so I I I recently published
17:05
well so that so I I I recently published a Blog on this so I'm uh I would like to
17:08
a Blog on this so I'm uh I would like to
17:08
a Blog on this so I'm uh I would like to share that with the listeners later but
17:11
share that with the listeners later but
17:11
share that with the listeners later but so so yeah I mean like just like Paul
17:12
so so yeah I mean like just like Paul
17:12
so so yeah I mean like just like Paul said right starting with those
17:13
said right starting with those
17:13
said right starting with those principles early on thinking about all
17:17
principles early on thinking about all
17:17
principles early on thinking about all these privacy transparency
17:18
these privacy transparency
17:18
these privacy transparency accountability there are six principles
17:21
accountability there are six principles
17:21
accountability there are six principles that Microsoft recommends and thinking
17:23
that Microsoft recommends and thinking
17:23
that Microsoft recommends and thinking about the goals of your use case through
17:26
about the goals of your use case through
17:26
about the goals of your use case through the lens of these principles then as
17:29
the lens of these principles then as
17:29
the lens of these principles then as lishing an REI strategy through people
17:31
lishing an REI strategy through people
17:31
lishing an REI strategy through people process and Technology when I say people
17:33
process and Technology when I say people
17:33
process and Technology when I say people like having strong accountability
17:35
like having strong accountability
17:35
like having strong accountability systems a governance system like
17:38
systems a governance system like
17:38
systems a governance system like research policy engineering teams
17:39
research policy engineering teams
17:39
research policy engineering teams working together and then from the
17:41
working together and then from the
17:41
working together and then from the process standpoint having something like
17:43
process standpoint having something like
17:43
process standpoint having something like red teaming in your application uh
17:45
red teaming in your application uh
17:45
red teaming in your application uh development process so red teaming is
17:48
development process so red teaming is
17:48
development process so red teaming is proactively probing for risks in your
17:50
proactively probing for risks in your
17:50
proactively probing for risks in your application making sure uh they don't
17:54
application making sure uh they don't
17:54
application making sure uh they don't behave negatively right and then also
17:57
behave negatively right and then also
17:57
behave negatively right and then also designing metric RS U for risk and then
18:01
designing metric RS U for risk and then
18:01
designing metric RS U for risk and then measuring them at scale and using some
18:03
measuring them at scale and using some
18:03
measuring them at scale and using some tools like Azure content safety to
18:05
tools like Azure content safety to
18:05
tools like Azure content safety to mitigate those risks and then at the
18:07
mitigate those risks and then at the
18:07
mitigate those risks and then at the technology level uh like thinking about
18:10
technology level uh like thinking about
18:10
technology level uh like thinking about risk mitigation at every layer of
18:12
risk mitigation at every layer of
18:12
risk mitigation at every layer of Technology stack uh like one of them is
18:16
Technology stack uh like one of them is
18:16
Technology stack uh like one of them is having meta proms it's like a simple
18:18
having meta proms it's like a simple
18:18
having meta proms it's like a simple solution having like good well-defined
18:20
solution having like good well-defined
18:20
solution having like good well-defined meta proms that can you know improve the
18:23
meta proms that can you know improve the
18:23
meta proms that can you know improve the behavior of your models so things like
18:26
behavior of your models so things like
18:26
behavior of your models so things like that and also making sure you select
18:28
that and also making sure you select
18:28
that and also making sure you select models that adher to fairness and low
18:32
models that adher to fairness and low
18:32
models that adher to fairness and low toxicity and things like that there are
18:34
toxicity and things like that there are
18:34
toxicity and things like that there are various benchmarks that can help you
18:36
various benchmarks that can help you
18:36
various benchmarks that can help you give you that
18:37
give you that
18:37
give you that information okay yeah yeah that makes
18:39
information okay yeah yeah that makes
18:39
information okay yeah yeah that makes sense so tell me have you um you don't
18:41
sense so tell me have you um you don't
18:41
sense so tell me have you um you don't have to mention any any brand names
18:43
have to mention any any brand names
18:43
have to mention any any brand names because you can't uh but but have you
18:45
because you can't uh but but have you
18:46
because you can't uh but but have you found any uh customer case where they
18:48
found any uh customer case where they
18:48
found any uh customer case where they they discovered that they had a lot of
18:50
they discovered that they had a lot of
18:51
they discovered that they had a lot of data maybe but their data was there was
18:53
data maybe but their data was there was
18:53
data maybe but their data was there was bias in their data they had challenges
18:56
bias in their data they had challenges
18:56
bias in their data they had challenges with that have has that happened like is
18:57
with that have has that happened like is
18:57
with that have has that happened like is that a thing that you look at your own
19:00
that a thing that you look at your own
19:00
that a thing that you look at your own data and you just realize that oh wow I
19:02
data and you just realize that oh wow I
19:02
data and you just realize that oh wow I didn't know that that was in
19:05
didn't know that that was in
19:05
didn't know that that was in there yeah there is um primarily we see
19:08
there yeah there is um primarily we see
19:08
there yeah there is um primarily we see it in HR areas as well I won't go into
19:11
it in HR areas as well I won't go into
19:11
it in HR areas as well I won't go into details but no no no you can't go into
19:13
details but no no no you can't go into
19:13
details but no no no you can't go into details I'm not asking you but but but
19:15
details I'm not asking you but but but
19:15
details I'm not asking you but but but but effectively that right when you once
19:17
but effectively that right when you once
19:17
but effectively that right when you once you crunch all your data and and you
19:20
you crunch all your data and and you
19:20
you crunch all your data and and you start asking it questions y a bunch of
19:22
start asking it questions y a bunch of
19:22
start asking it questions y a bunch of really weird things come out and it's
19:24
really weird things come out and it's
19:24
really weird things come out and it's like what is that in my data yeah some
19:27
like what is that in my data yeah some
19:27
like what is that in my data yeah some of it is am yeah you know data is data
19:30
of it is am yeah you know data is data
19:30
of it is am yeah you know data is data and how you interpret is is just another
19:32
and how you interpret is is just another
19:32
and how you interpret is is just another layer yeah exactly yeah yeah that's
19:34
layer yeah exactly yeah yeah that's
19:35
layer yeah exactly yeah yeah that's interesting having that responsible AI
19:36
interesting having that responsible AI
19:36
interesting having that responsible AI layer is critical in every every
19:39
layer is critical in every every
19:39
layer is critical in every every deployment yes so so real like as a
19:42
deployment yes so so real like as a
19:42
deployment yes so so real like as a final thing here for this this quick
19:44
final thing here for this this quick
19:44
final thing here for this this quick conversation today um like can you give
19:46
conversation today um like can you give
19:46
conversation today um like can you give some advice on on like starting the
19:48
some advice on on like starting the
19:48
some advice on on like starting the cloud uh an AI generative AI Journey
19:51
cloud uh an AI generative AI Journey
19:51
cloud uh an AI generative AI Journey like what what does a company do like as
19:54
like what what does a company do like as
19:54
like what what does a company do like as as concise as possible if you can but
19:56
as concise as possible if you can but
19:56
as concise as possible if you can but you know what I mean
19:58
you know what I mean
19:59
you know what I mean go
19:59
go
19:59
go ah yeah so my advice would be to start
20:03
ah yeah so my advice would be to start
20:03
ah yeah so my advice would be to start with something like an ideation phase
20:05
with something like an ideation phase
20:05
with something like an ideation phase which I told you right like if they are
20:06
which I told you right like if they are
20:06
which I told you right like if they are starting now in generative and like
20:08
starting now in generative and like
20:08
starting now in generative and like doing hackathons which are a good
20:10
doing hackathons which are a good
20:10
doing hackathons which are a good breeding ground for like creative
20:11
breeding ground for like creative
20:11
breeding ground for like creative innovative solutions and then yep and
20:14
innovative solutions and then yep and
20:14
innovative solutions and then yep and then picking a lwh hanging fruit use
20:16
then picking a lwh hanging fruit use
20:16
then picking a lwh hanging fruit use case which could be like a simple
20:18
case which could be like a simple
20:18
case which could be like a simple internal co-pilot like an organizational
20:21
internal co-pilot like an organizational
20:21
internal co-pilot like an organizational chatbot a Rags solution right using
20:24
chatbot a Rags solution right using
20:24
chatbot a Rags solution right using Azure opener on your data something like
20:26
Azure opener on your data something like
20:26
Azure opener on your data something like that which will help you uh get to quick
20:28
that which will help you uh get to quick
20:28
that which will help you uh get to quick wins and and show value as well so and
20:32
wins and and show value as well so and
20:32
wins and and show value as well so and also get your hands dirty and your
20:34
also get your hands dirty and your
20:34
also get your hands dirty and your employees can learn about generative so
20:36
employees can learn about generative so
20:36
employees can learn about generative so that's the first thing and the second
20:38
that's the first thing and the second
20:38
that's the first thing and the second thing I would say is like when you're
20:39
thing I would say is like when you're
20:39
thing I would say is like when you're going to production thinking about
20:41
going to production thinking about
20:41
going to production thinking about latency and ux is very important that's
20:44
latency and ux is very important that's
20:44
latency and ux is very important that's what I have realized because uh if if uh
20:48
what I have realized because uh if if uh
20:48
what I have realized because uh if if uh it just takes one or two bad experiences
20:50
it just takes one or two bad experiences
20:50
it just takes one or two bad experiences for your users to switch the platform so
20:54
for your users to switch the platform so
20:54
for your users to switch the platform so latency and a good well-designed ux I
20:56
latency and a good well-designed ux I
20:56
latency and a good well-designed ux I think is our critical when you're going
20:58
think is our critical when you're going
20:58
think is our critical when you're going to
20:59
to
20:59
to production that makes sense so
21:01
production that makes sense so
21:01
production that makes sense so effectively you're saying like it's it's
21:03
effectively you're saying like it's it's
21:03
effectively you're saying like it's it's a step stepwise right start hacking get
21:06
a step stepwise right start hacking get
21:07
a step stepwise right start hacking get something anything that works U and and
21:09
something anything that works U and and
21:09
something anything that works U and and just feel it and like okay how did how
21:12
just feel it and like okay how did how
21:12
just feel it and like okay how did how did this feel like what is what is this
21:14
did this feel like what is what is this
21:14
did this feel like what is what is this about and then take the next step on top
21:16
about and then take the next step on top
21:16
about and then take the next step on top of that and as you mentioned Magnus it's
21:19
of that and as you mentioned Magnus it's
21:19
of that and as you mentioned Magnus it's a quick moving field like it's
21:21
a quick moving field like it's
21:21
a quick moving field like it's constantly changing so reading Vlogs or
21:23
constantly changing so reading Vlogs or
21:23
constantly changing so reading Vlogs or watching videos is also a great way to
21:25
watching videos is also a great way to
21:25
watching videos is also a great way to educate yourself fantastic well thank
21:28
educate yourself fantastic well thank
21:28
educate yourself fantastic well thank you you guys for being on the show today
21:29
you you guys for being on the show today
21:29
you you guys for being on the show today I I really appreciated this conversation
21:31
I I really appreciated this conversation
21:31
I I really appreciated this conversation and I think it clarified and and give
21:33
and I think it clarified and and give
21:33
and I think it clarified and and give some strategy and advice to a lot of of
21:35
some strategy and advice to a lot of of
21:35
some strategy and advice to a lot of of our our listeners so thank you guys for
21:38
our our listeners so thank you guys for
21:38
our our listeners so thank you guys for being on the cloud show thanks for
21:39
being on the cloud show thanks for
21:39
being on the cloud show thanks for having
21:40
having
21:40
having us yeah and uh for the audience I'll see
21:43
us yeah and uh for the audience I'll see
21:43
us yeah and uh for the audience I'll see you next time on the cloud show
21:46
you next time on the cloud show
21:46
you next time on the cloud show [Music]
#Computer Education
#Data Management