October 17, 2024 • 29 mins
Are you ready to take your AI career to the next level? We dive deep into the world of strategic partnerships, uncovering everything from finding the perfect match to harnessing the power of developer communities. Get ready for insider tips that will help you build the future of AI.
We explore the game-changing potential of AI partnerships—how can businesses and developers come together to create groundbreaking solutions? What’s the secret sauce to a successful collaboration? We also dive into the crucial role that developer communities and events play in driving innovation and building connections.
Join us as we explore these ideas with:
Jason Corso, Cofounder and Chief Science Officer, Voxel51
Paula Ramos, AI Evangelist, Intel
Christina Cardoza, Editorial Director, insight.tech
Jason and Paula answer our questions about
- The evolving artificial intelligence landscape
- How developers can keep up with the changes
- Gaining developer support from large companies
- Being part of developer communities and events
- Staying on top of upcoming AI trends
- Fostering community engagement
Related Content
For the latest innovations from Voxel, follow them on Twitter at @voxel51, LinkedIn, and GitHub. For the latest innovations from Intel, follow them on Twitter at @intel, LinkedIn, and GitHub.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
(bright music)
- Hello and welcome to"insight.tech Talk,"
where we explore the latest IoT, edge, AI,
and network technologytrends and innovations.
I'm your host, Christina Cardoza,
(00:21):
Editorial Director of insight.tech,
and today we're going to betalking about AI partnerships
that spark developerengagement and innovations.
Who better to discuss this withthan two companies embedded
in the AI and developer communities.
Today we'll be speakingto Paula Ramos from Intel
as well as Jason Corso from Voxel51.
But as always, before we get started,
(00:42):
let's get to know our guests.
Paula, a good friend of the show.
For those of us who haven't heard
your previous conversations,
what can you tell us about yourself
and what you're doing at Intel?
- Yes, for sure.
Thank you, Christina, for having me here.
So I'm so excited. So I Paula Ramos.
I have a PhD in computervision and machine learning,
(01:02):
and I'm working at Intel as AI Evangelist,
working with multiple products
and multiple developers around the globe.
- Great, and Jason Corso from Voxel51.
First-time guest of the podcast.
What can you tell usabout yourself and Voxel?
- Likewise. Nice to meet you all.
Thanks for the invitation, Christina.
So Jason Corso.
Yeah, I have a PhD in computer science.
(01:25):
I'm a Co-Founder at Voxel.
At Voxel, we make a softwarerefinery to help you work
with your data, yourmodels, various needs,
and kind of refine theminto production visual AI.
I'm also on the facultyof robotics and EECS
at the University of Michigan,
where I've done researchfor the last 10 or 15 years
(01:48):
in computer vision and machine learning,
all at the boundariesbetween the physical world
and what we can do withcomputational systems these days.
- Awesome, so you've been inthis space for a long time now
and have probably seen it evolve even...
It feels like every daysomething new is happening,
and it's evolving even further.
So that's where I wanted to start off
the conversation with you, Jason.
(02:09):
If you could just talkabout what you're seeing
in this space, how it haschanged over the last few years,
where we are today
and what are the trendsshaping where we are.
- Yeah, indeed, it has changedquite a bit in the last,
even the last few years, alsothe last 20 years, right?
Like when I was doing my PhD,
we were looking at things about,
(02:30):
how you can use computervision to understand gestures
and so on to interact with the computer,
and look where we are now, right?
20 years later, it'sbeen quite a wild ride.
So last few years, let's see.
I think there probablyare two major developments
I would argue, right,in the last couple years
that really are driving theway we all think about AI.
(02:51):
So the first one is probablypretty obvious, right?
Like the availability ofthese large language models
that capture huge token lengths
and can embed actuallynatural human language
into the language modelthat's there to really give us
a resource in which we caninteract with rather naturally.
(03:12):
Now, I mean, there are anawful lot of questions around
what their limitations areand their capabilities are,
but at the same time,I think you'd be easy
to find lots of differentapplications, right?
Like, I think in thebeginning of this year,
I wrote some quick note on LinkedIn
about how I think LMs will evolve
in 2024 and this year.
(03:32):
One of those key elementsthat I thought was
that we would see like a true revolution
in how we think about search, right?
And just information search,information gathering,
and all that and so on.
And I think we really arebeginning to see that.
I think on the other one, though,
I'd probably point to an appreciation
for the role that data has begun to play
(03:55):
or has been playing in the development
of various AI ML models, right?
Everyone when you go toschool, in grad school, right,
like you take yourmachine learning course,
and you go and start training models
to recognize digits and so on,
you just go quickly download a dataset.
Either it's from some repository
or your professor gives it to you,
and most of the focus is on the algorithm.
(04:16):
And so we've built this cultureof like the model is king,
but if you really think aboutwhat's happening, right,
even like variousleaders in the LLM space,
to bring back to the first one,
are have begun to talkabout the critical role
that data, good data,high quality data plays
in this marriage of model, code, and data
(04:38):
to build the AI systems that we're using.
So I don't know exactly where
that appreciation isgoing to lead us, right?
Like at my company, forexample, we focus heavily
on the role that data playsand providing developer tools
for engaging with dataalongside their models
rather than just expectingyou to gen up some scripts
(05:00):
to visualize your data or whatever, right?
But I think it's good for me
because it's a long time since when I was,
like 20 years ago, my datasetswere dozens of samples,
hundreds of samples, right?
Now we have data sets that are dozens
of millions of samples or whatever, right?
So actually managing them
and understanding the failure modes
and the distribution andso on is very difficult
(05:22):
and requires, I think, new thinking.
- Yeah, absolutely.
And you mentioned the searchand information gathering.
I'm definitely seeing on the consumer side
AI being more prominent in these areas.
When I search on Google or anything now,
instead of just getting a list of links,
an answer from from Gemini comes up.
(05:43):
So it's interesting tosee how AI is evolving,
but I'm glad you broughtup LLMs, the repositories,
and algorithms, and thisdata, and these models
because it's really the developers
that are pushing theseadvancements forward.
A lot of times on "insight.tech,"
we're writing about advancements
in manufacturing and retail and education,
how businesses are using AIto transform their spaces,
(06:04):
but what's behind these transformations
are really developers thatare building these solutions
that are working with LLMs.
So Paula, I'm curious from your take,
'cause you work with a lot of developers,
you talk to a lot ofdevelopers in this space,
what has their role been
in keeping up with AI, andhow can they even continue
(06:27):
to compete in this spacewith all of the advancements
and skillsets happening?
- Yeah, that is a great question.
I think that all of thedevelopers are looking
for their path every day
because the things are changing so fast.
But the main things thatwe need to have in mind
as the developers, whatkind of challenges we have,
is that we need to drive innovation
in a huge field that is there: artificial intelligence. (06:49):
undefined
So we need to be creative,
we need to buildintelligence applications,
and we need to solve problems.
So maybe we have the same problems
that we have 20 years ago,as Jason was mentioning,
but we have better tools right now.
We have a better way toapproach those solutions,
but we need to be so creativewith those solutions.
(07:12):
So still we have a lot oftools, and we need to think
about the final user of the applications.
So I think that there aresome challenges right now
in still we have room to improve,
that is model development,data management,
or how we can deploy thosemodels in the easy way.
So we need to use a cloud system
(07:34):
or we can use an edge solution.
So we need to think aboutindependent of that, for sure.
The skills that we need tofind could be different,
but basically havingdevelopers programming
in different kind of languages,
organizing or producingdifferent kind of datasets.
(07:56):
Also something that is really important
in this field is theopen-source community.
Open-source community ischanging the cadence of the AI
because when we have thesemodels open to everyone,
they can access those models,
so they can access those datasets,
and improve and improvethose models round by round,
(08:17):
of those datasets round by round.
So I think that theresponsibility that we have
as a developers is hugein this new era of AI.
For sure, I think roles arein different kind of sectors.
We can talk about manufacturing,retail, but more than that
is what kind of problemwe want to solve today.
(08:40):
Could be complex, could be simple,
but the solution always willbe the simple as possible,
and this is the main challenge
that developers have right now.
- Yeah, I love how you saidwe need to drive innovations,
we need to createintelligent applications,
we need to solve problems,
because developers aren't in it alone.
(09:04):
They don't have to build it from scratch.
They can leverage partnerslike Intel and Voxel
and community members tomake some of this happen.
For instance, I love that Intelhas the Edge Reference Kits,
and sometimes you guys are walking them
through how to build a solution
and giving them thecode to do self-checkout
or to build something in manufacturing,
(09:25):
and they can just customizeit after they learn
a little bit more aboutit and how to do that.
So I'm curious, in what otherways can developers partner
with companies like Intel andhow that's going to benefit them
to reach out into these differentareas and to ask for help
or ask questions and be apart of the Intel community
(09:48):
or other open-source communities?
- That's a great question.
So we have multiple channels right now.
As you mentioned, we havethe Edge Reference Kits
that developers can access.
In an easy way, they can find a solution,
complex problem with an easy solution,
where we are trying to showthem with tutorials, code,
(10:08):
videos how they can navigatethat specific vertical,
manufacturing, retail,healthcare, LLMs as well,
and working with multiple models.
So Intel has a variety of solutions.
So basically we have solutions.
We have hardwareaccelerators for retraining,
for fine tuning models,
(10:28):
but also we have solutionsthat work at the edge.
Or also you can use your laptop.
So you can use yourlaptops to work with AI.
And we are creating a specific framework
that is called OpenVINO,
where developers can use OpenVINO
(10:49):
to optimize and quantize model.
That means that they can use
the same infrastructure that they have,
they can use the same computer,
and they can run LLMs,
optimize and quantize LLMs,INTEGER*4, for example,
or they can use the integrated GPU
that Intel also provide the processors.
So I think that Intel
(11:11):
with OpenVINO is enabling developers
to easily prove and test these LLMs.
And this is just one stepbehind the real solution,
the solution that we want toput in the production systems
so they can createpilots, they can impress
(11:31):
the bosses with the tutorials and examples
that they can run in theirown laptops before to move
to the real or thefinal production system.
And Intel has this possibility also.
Developers can accessIntel Developer Cloud
to test multiple hardwarebefore to buy that hardware.
(11:53):
That is really cool, andalso accessing accelerators
and accessing also thelatest, for example, AI PC.
So we are provisioning alot of tools to developers,
and also we have, I almost miss that,
but we have an amazing repository
where developers can testthe latest AI trends.
(12:15):
So we have OpenVINO notebook's repository,
where if something happenedtoday, literally in two days,
we will see the notebook withthat specific model for sure.
This is for the open-source community.
So you can test there,for example, Llama 3.1,
YOLOv10, and the latest AI trends.
(12:36):
And this is a great tool.
And the most important thing is
we are not forcing developersto buy specific hardware
to run those models as well.
So developers can also run these models
in the actual hardware that they have.
We are also supporting ARM,
and we are also supportinga variety of Intel hardware.
(12:58):
Also integrated GPUs,
that is the most usage,
an integrated GPU, thatwe can see in the world.
- Yeah, it's great thatyou are making it easy
for developers to getstarted with the equipment
or hardware that they have.
And a lot of the kits and the challenges
we were just talkingabout, and repositories,
these are ongoing thingsthat are available
(13:20):
to developers at any time.
But I'm thinking about, I know recently,
which probably feels likeforever ago, you were at CVPR,
and there was a competitionand challenging going there.
So that's more of aone-off timely challenge
that is available sometimesto developers going
to these events, havingthese things happen.
(13:41):
So I'm curious, Jason, 'causeI know the company was also
at that event, but there'sother events that you guys host
or that you're at that havethese developer challenges.
I'm curious, what wouldyou say is the importance
of developers going to these events,
engaging in these communities,
and participating in someof those competitions?
- Yeah, it's a good point, right?
(14:02):
I mean, like even just before CVPR,
Voxel had our first in-person hackathon,
actually in New York City,
and it's that type ofengagement where we really see
like excited developersengaging with new technology
and then really trying towork together on new teams
to kind of solve a new problem.
That was really fun, butI think like one key angle
(14:24):
for developer events isobviously education, right?
Like learning new things.
And I think if you take my earlier answer
about how AI has evolved
and think about like a keytrend for the future, right?
Like a key trend that we're seeing
for the future is languagecombined with vision
combined with new compute capabilities
and openly available dataand these foundational models
(14:47):
to really tackle new problems
in what at Voxel we call visual AI.
I think we're going to seeincreasing contributions
to that effect, but how do you do it?
What do you do, right?
One has to go to developer events
or other types of conferenceslike CVPR or whatever, truly,
to really stay abreast ofwhat's happening there.
I mean, for me it's, in some sense,
(15:07):
the educational sideis very natural, right?
I'm a faculty member. I teach.
I'm not teaching right now this year,
but last year I taughtintro to computer vision.
So like three hours a week I was doing
this developer event, in some sense,
for 300 students to learnabout computer vision.
So I think one thingwe've learned at Voxel
is this AI space is evolving so rapidly
(15:28):
that it seems like everyone,
even faculty members who'vebeen in the field for ages,
we're in like constantinformation-gathering mode.
It's impossible to stayup to date with everything
from like cutting-edgeresearch papers, on one hand,
all the way to what are the new APIs
and libraries that you have tothen, that you have to learn.
And so to do this, at least at Voxel,
(15:48):
what we've tried to do ismaintain like a weekly,
at least one per week,if not more per week,
sort of technical output thatin some form of an event,
like different formats,
that really allows thecommunity to stay engaged.
So we have an eventscalendar at voxelv51.com
that we can include in the show notes.
(16:08):
I think we have something liketwo dozen events scheduled
between now and the end of the year.
Just personally, for example,every Monday at noon Eastern,
I maintain these open office hours
where kind of anyone can sign into them.
They're on Zoom.
We talk everything from like...
A couple weeks ago we werereviewing someone's paper,
(16:29):
and we went through slides
and like actual like technicalmodel all the way to like...
Oftentimes I get asked like,"This is my first time thinking
about getting into computer vision.
What should I look at first?"
Right? So pretty broad.
But we have some hackathons,virtual meetups, and so on.
So I think that it's like raw education
just about foundational capabilities,
(16:51):
but also these developerevents really help engagement
just from like a staying upto date with what's happening.
- That's great, and that's awesome
that you have those open hours
that developers can justjoin and start to learn.
I'm curious because obviouslythere are virtual conferences,
then there can be conferencesin different parts
(17:12):
of the world, and it canbe tough for developers.
They can't go to all of them
or there's just so many outthere, it's hard to choose from.
Is there anything coming upthat you want to call out
that developers shouldhave on their radar?
Or is there anything,
any other resourcesavailable to them online
(17:33):
that you think that theyshould take advantage of?
- What Paula was sayingearlier, being open-source
is like the gateway tofostering innovation, right?
Like our software atVoxel51 is called FiftyOne.
It's on GitHub.
We have the permissive licensing
for the open-source component of it,
which is basically like oneuser, one machine local data.
(17:56):
You can fork it. You can submit PRs.
We make releases.
I think it's on the orderof every one to two months.
Every release that wehave has some content
from our community,and we've been educated
so much over the last fouryears since we released it
about from community needsand community contributions.
Most recently we have this newfunctionality called Panels,
(18:20):
which FiftyOne is basicallylike a visual component
as well as a softwareSDK for doing the work
that we're talking about here,
like data and model refinement,
but with Panels youcan build functionality
for the front end withoutknowing how to write React
or JavaScript or anything with UX.
You can write it right inPython, and all of a sudden,
you can still enhancethe GUI functionality.
(18:42):
So I think those are great ways,
actual events but alsojust becoming a part
of open-source projects is another way
to really to get involved inthe developer ecosystem for AI.
- Yeah, absolutely, andI think it also helps,
companies like yourself whohave these open-source models.
You might not have picked up on something
that somebody in the developercommunity picks up on,
(19:03):
and they can really bea part of that community
and make changes and pointthings out and contribute
to companies and projects like yourself.
So it's always great to bea part of those discussions,
see what's going on,hearing what developers
are talking about as well assome of the ongoing challenges
(19:23):
that they're facing in these spaces.
Paula, I know OpenVINO,
there's a huge GitHubcommunity around there,
and you mentioned a little bit of the kits
and some other things thatIntel offers, but I'm curious,
in what other ways doesIntel foster that innovation
and that communityengagement for developers?
(19:44):
- That is a great question
because we have been workingso hard on that part as well.
So we have multiple ways.
So we are creating multiple ways
to create this innovation with developers.
So we have one program, thatis the Innovator Program,
where we have multipledevelopers around the globe,
that they can try, theycan test technology.
(20:05):
They can make their own applications
and they can share that with us.
So just stay tuned, for example,
in my LinkedIn or in my network as well.
So we are highlightingsome of these innovators.
This is one thing that we have.
And basically they createtheir own repository.
They fork their repository andthey create new applications
or improve the applicationwith the contribution.
(20:29):
So another thing that we haveis Google Summer of Code.
We have a program with Google every year
where we have multiple proposals
and we have several developers
around the globe as wellworking with us for three months
with different mentorsin the OpenVINO team.
And, for example, youmentioned about CVPR.
(20:51):
So we worked with Anomalib.
There is a library that also we have
in the OpenVINO ecosystem.
And we have two proposalslast year about Anomalib,
and one of these proposals,
the student that was involvedin Google Summer of Code,
and the mentors and the professor as well,
they create a paper.
(21:13):
The paper was submitted in the workshop
of the anomaly detection,
Visual Anomaly InspectionWorkshop at CVPR,
and that was accepted.
So we are closing also the gap
in between industry and theacademia with conferences.
So we are also participatingwith the students,
developers in those conferencesthrough programs as,
(21:35):
for example, Google Summer ofCode, but no more than that.
For sure, we are moving so fast also
in relations with universities,
what kind of things we canwork with universities,
helping them to create some research
and research proposals thatIntel also can support.
(21:55):
At CVPR, we are sponsoring as well
the challenge in this workshopabout anomaly detection.
We try also to in invite developers,
and we create a marketing campaign
around the challenge to invite developers
to participate in that challenge.
We receive more than 400 participants
(22:19):
and more than 100 submissions.
That was an amazing and remarkable number
around maybe one month anda half that we received,
and we can see how the knowledge is moving
in using anomaly detection.
For sure, talking about OpenVINO,we have multiple things.
As I mentioned before, OpenVINOis an open-source tool,
(22:42):
and we have a repository
when we have differentkind of contributions
depending of the product.
So we have OpenVINO, OpenVINO notebooks.
We have OpenVINO build and deploy.
In that repository,OpenVINO build and deploy,
You can find all the Edge Reference Kits
that we have been talking about today.
(23:04):
OpenVINO notebooks, youcan find the tutorial,
and in the OpenVINO repository,you can find the API.
So we have a huge ecosystem
where we are trying to touchnot just the inference part,
also the training part withanomaly detection, Anomalib,
and also OpenVINO Training Extension.
So we have a huge ecosystem
(23:25):
that I really want toinvite all the developers
and all the people thatare watching this podcast
or listening this to this podcast
to visit those repositories,
visit the organization,"openvinotoolkit" in GitHub,
and you can find all the repositories
that I'm talking about.
- Absolutely.
It's exciting hearing all ofthese different resources,
all these different waysdevelopers can get started.
(23:48):
So I'm excited to see movingforward what types of solutions
and innovations developerscontinue to build,
and I hope they take you guysup on some of these events
and meet you, whether that'sin person or virtually.
I know sometimes it can be intimidating
when you're gettingstarted in these areas,
but having companies like Voxel
and Intel support developers,that's great to see.
(24:11):
And I also saw, Jason,
in addition to the virtual office hours,
there's availability todo one-on-one meetings.
So if developers feel intimidated somehow
or don't want to ask aquestion in a group setting,
it's great that you guys aremaking yourself available
to help developers whenand where they need it.
(24:31):
So I want to thank you both again
for joining us on this podcast.
Before we go, if there'sany final thoughts
or key takeaways you wantto leave developers with
as they go on this journey,engage with each other,
and engage with yourselves.
Jason, I'll start with you.
- Great. Yeah, thanks very much.
So I mean, first parting thought would be
(24:52):
that I think I just wantto express my thanks
to the developercommunity that we've built
over the last four or five years, right?
We wouldn't be where we aretoday without the community.
It's such a vibrant and rich environment.
But like second thing isthat actually we're hiring.
So we're hiring developers.
(25:12):
I mean, actually across theboard we are as we grow,
after we closed our SeriesB earlier this year,
but for this conversation,machine learning engineer roles,
both for core engineering work
as well as developerrelations work, right?
We believe in developers so much,
so we hire individualsthat are fully trained
and can write papers, canwrite code, and so on,
(25:33):
but their role isactually building bridges
with the community.
And then maybe just the lastparting remark is that we...
As a company, we are open-source driven,
but we do actually havedozens of customers
that use our commercial enterprise version
that we call FiftyOne Teams.
It kind of like relaxes thatindividual user local data work
(25:55):
and allows you to develop thesame functionality together
in teams, in the cloud or on-prem.
And we love to engage inconversations around FiftyOne Teams
as well with with your community.
We have customers, many ofwhich are in the Fortune 500,
but across manufacturing,security, automotive, right?
(26:15):
Like pretty broad basecustomer base. So thanks.
- Yeah, absolutely love tohear about job openings.
It shows this space is growing,
this space is becoming important,
and some of the innovationsand transformations
that we talk about on "insight.tech"
wouldn't be possible without developers.
So exciting opportunityfor anybody listening
(26:38):
to go join the Voxel51 team.
Paula, always love havingyou on the podcast.
Thank you, again.
I feel like every conversation,
there's something new to talk about,
something new happening in the AI space,
so curious what our nextconversation will be about.
But before we go, arethere any final thoughts
or key takeaways youwant to leave with us?
- Yes, for sure. Sofirst of all, thank you.
(26:59):
Thank you, Christina,for create this space
to talk about what we have,
and thank you also to Voxel51.
We have been creating agreat relation with Voxel51.
Different conferences, we tryto share some space together.
And this also talks pretty well
about that we have the real intention
(27:21):
to work in the open-source community.
So we are open to work with all of you,
try to find the best path to developers,
because here the mostimportant thing are developers.
So the company for sureis really important.
We have a lot of thingsto learn from the company,
what kind of products we can provide,
what kind of tools we canprovide to developers.
(27:43):
And always we are thinking that, I mean,
we need to enable you to usethis hardware in software
that we can provideand you can accelerate,
you can improve yourpipelines and your workloads.
That is the main intention.
We have right now a lot ofthings to share with you.
So we talk about OpenVINO,Edge Reference Kits,
(28:05):
but more things are coming in the future.
For example, we have the new,
you have the new AI PC that you can try.
We have a new enginein the microprocessor,
that is the NPU, Neural Processing Unit,
that we can also expediteand accelerate part
of the conventional and generative AI,
conventional AI, generative AI,
(28:27):
process with that small device.
This is one of the thingsthat we can talk about
(chuckling) in the future,Christina, for sure.
Thank you again,
and I'm looking forward toconnect with all of you.
- Absolutely, and you talked about earlier
how some of these innovations
or these tools you haveavailable is making it easy
for developers to start working
(28:49):
no matter what hardware they're using,
and the AI PC just makesit that much easier
for the AI development,deployment, performance
of your solutions, all that great stuff.
So I know Intel has a lotof resources around AI PCs
that we'll make sure toprovide to developers as well.
But thank you both againfor joining us today.
Thank you to Intel and Voxel51for these great resources
(29:11):
and communities you'vecreated for developers
and spaces for them to getstarted and get that support.
Until next time, this hasbeen "insight.tech Talk."
(bright music)
- Hello and welcome to"insight.tech Talk,"
where we explore the latest IoT, edge, AI,
and network technologytrends and innovations.
I'm your host, Christina Cardoza,
(00:21):
Editorial Director of insight.tech,
and today we're going to betalking about AI partnerships
that spark developerengagement and innovations.
Who better to discuss this withthan two companies embedded
in the AI and developer communities.
Today we'll be speakingto Paula Ramos from Intel
as well as Jason Corso from Voxel51.
But as always, before we get started,
(00:42):
let's get to know our guests.
Paula, a good friend of the show.
For those of us who haven't heard
your previous conversations,
what can you tell us about yourself
and what you're doing at Intel?
- Yes, for sure.
Thank you, Christina, for having me here.
So I'm so excited. So I Paula Ramos.
I have a PhD in computervision and machine learning,
(01:02):
and I'm working at Intel as AI Evangelist,
working with multiple products
and multiple developers around the globe.
- Great, and Jason Corso from Voxel51.
First-time guest of the podcast.
What can you tell usabout yourself and Voxel?
- Likewise. Nice to meet you all.
Thanks for the invitation, Christina.
So Jason Corso.
Yeah, I have a PhD in computer science.
(01:25):
I'm a Co-Founder at Voxel.
At Voxel, we make a softwarerefinery to help you work
with your data, yourmodels, various needs,
and kind of refine theminto production visual AI.
I'm also on the facultyof robotics and EECS
at the University of Michigan,
where I've done researchfor the last 10 or 15 years
(01:48):
in computer vision and machine learning,
all at the boundariesbetween the physical world
and what we can do withcomputational systems these days.
- Awesome, so you've been inthis space for a long time now
and have probably seen it evolve even...
It feels like every daysomething new is happening,
and it's evolving even further.
So that's where I wanted to start off
the conversation with you, Jason.
(02:09):
If you could just talkabout what you're seeing
in this space, how it haschanged over the last few years,
where we are today
and what are the trendsshaping where we are.
- Yeah, indeed, it has changedquite a bit in the last,
even the last few years, alsothe last 20 years, right?
Like when I was doing my PhD,
we were looking at things about,
(02:30):
how you can use computervision to understand gestures
and so on to interact with the computer,
and look where we are now, right?
20 years later, it'sbeen quite a wild ride.
So last few years, let's see.
I think there probablyare two major developments
I would argue, right,in the last couple years
that really are driving theway we all think about AI.
(02:51):
So the first one is probablypretty obvious, right?
Like the availability ofthese large language models
that capture huge token lengths
and can embed actuallynatural human language
into the language modelthat's there to really give us
a resource in which we caninteract with rather naturally.
(03:12):
Now, I mean, there are anawful lot of questions around
what their limitations areand their capabilities are,
but at the same time,I think you'd be easy
to find lots of differentapplications, right?
Like, I think in thebeginning of this year,
I wrote some quick note on LinkedIn
about how I think LMs will evolve
in 2024 and this year.
(03:32):
One of those key elementsthat I thought was
that we would see like a true revolution
in how we think about search, right?
And just information search,information gathering,
and all that and so on.
And I think we really arebeginning to see that.
I think on the other one, though,
I'd probably point to an appreciation
for the role that data has begun to play
(03:55):
or has been playing in the development
of various AI ML models, right?
Everyone when you go toschool, in grad school, right,
like you take yourmachine learning course,
and you go and start training models
to recognize digits and so on,
you just go quickly download a dataset.
Either it's from some repository
or your professor gives it to you,
and most of the focus is on the algorithm.
(04:16):
And so we've built this cultureof like the model is king,
but if you really think aboutwhat's happening, right,
even like variousleaders in the LLM space,
to bring back to the first one,
are have begun to talkabout the critical role
that data, good data,high quality data plays
in this marriage of model, code, and data
(04:38):
to build the AI systems that we're using.
So I don't know exactly where
that appreciation isgoing to lead us, right?
Like at my company, forexample, we focus heavily
on the role that data playsand providing developer tools
for engaging with dataalongside their models
rather than just expectingyou to gen up some scripts
(05:00):
to visualize your data or whatever, right?
But I think it's good for me
because it's a long time since when I was,
like 20 years ago, my datasetswere dozens of samples,
hundreds of samples, right?
Now we have data sets that are dozens
of millions of samples or whatever, right?
So actually managing them
and understanding the failure modes
and the distribution andso on is very difficult
(05:22):
and requires, I think, new thinking.
- Yeah, absolutely.
And you mentioned the searchand information gathering.
I'm definitely seeing on the consumer side
AI being more prominent in these areas.
When I search on Google or anything now,
instead of just getting a list of links,
an answer from from Gemini comes up.
(05:43):
So it's interesting tosee how AI is evolving,
but I'm glad you broughtup LLMs, the repositories,
and algorithms, and thisdata, and these models
because it's really the developers
that are pushing theseadvancements forward.
A lot of times on "insight.tech,"
we're writing about advancements
in manufacturing and retail and education,
how businesses are using AIto transform their spaces,
(06:04):
but what's behind these transformations
are really developers thatare building these solutions
that are working with LLMs.
So Paula, I'm curious from your take,
'cause you work with a lot of developers,
you talk to a lot ofdevelopers in this space,
what has their role been
in keeping up with AI, andhow can they even continue
(06:27):
to compete in this spacewith all of the advancements
and skillsets happening?
- Yeah, that is a great question.
I think that all of thedevelopers are looking
for their path every day
because the things are changing so fast.
But the main things thatwe need to have in mind
as the developers, whatkind of challenges we have,
is that we need to drive innovation
in a huge field that is there: artificial intelligence. (06:49):
undefined
So we need to be creative,
we need to buildintelligence applications,
and we need to solve problems.
So maybe we have the same problems
that we have 20 years ago,as Jason was mentioning,
but we have better tools right now.
We have a better way toapproach those solutions,
but we need to be so creativewith those solutions.
(07:12):
So still we have a lot oftools, and we need to think
about the final user of the applications.
So I think that there aresome challenges right now
in still we have room to improve,
that is model development,data management,
or how we can deploy thosemodels in the easy way.
So we need to use a cloud system
(07:34):
or we can use an edge solution.
So we need to think aboutindependent of that, for sure.
The skills that we need tofind could be different,
but basically havingdevelopers programming
in different kind of languages,
organizing or producingdifferent kind of datasets.
(07:56):
Also something that is really important
in this field is theopen-source community.
Open-source community ischanging the cadence of the AI
because when we have thesemodels open to everyone,
they can access those models,
so they can access those datasets,
and improve and improvethose models round by round,
(08:17):
of those datasets round by round.
So I think that theresponsibility that we have
as a developers is hugein this new era of AI.
For sure, I think roles arein different kind of sectors.
We can talk about manufacturing,retail, but more than that
is what kind of problemwe want to solve today.
(08:40):
Could be complex, could be simple,
but the solution always willbe the simple as possible,
and this is the main challenge
that developers have right now.
- Yeah, I love how you saidwe need to drive innovations,
we need to createintelligent applications,
we need to solve problems,
because developers aren't in it alone.
(09:04):
They don't have to build it from scratch.
They can leverage partnerslike Intel and Voxel
and community members tomake some of this happen.
For instance, I love that Intelhas the Edge Reference Kits,
and sometimes you guys are walking them
through how to build a solution
and giving them thecode to do self-checkout
or to build something in manufacturing,
(09:25):
and they can just customizeit after they learn
a little bit more aboutit and how to do that.
So I'm curious, in what otherways can developers partner
with companies like Intel andhow that's going to benefit them
to reach out into these differentareas and to ask for help
or ask questions and be apart of the Intel community
(09:48):
or other open-source communities?
- That's a great question.
So we have multiple channels right now.
As you mentioned, we havethe Edge Reference Kits
that developers can access.
In an easy way, they can find a solution,
complex problem with an easy solution,
where we are trying to showthem with tutorials, code,
(10:08):
videos how they can navigatethat specific vertical,
manufacturing, retail,healthcare, LLMs as well,
and working with multiple models.
So Intel has a variety of solutions.
So basically we have solutions.
We have hardwareaccelerators for retraining,
for fine tuning models,
(10:28):
but also we have solutionsthat work at the edge.
Or also you can use your laptop.
So you can use yourlaptops to work with AI.
And we are creating a specific framework
that is called OpenVINO,
where developers can use OpenVINO
(10:49):
to optimize and quantize model.
That means that they can use
the same infrastructure that they have,
they can use the same computer,
and they can run LLMs,
optimize and quantize LLMs,INTEGER*4, for example,
or they can use the integrated GPU
that Intel also provide the processors.
So I think that Intel
(11:11):
with OpenVINO is enabling developers
to easily prove and test these LLMs.
And this is just one stepbehind the real solution,
the solution that we want toput in the production systems
so they can createpilots, they can impress
(11:31):
the bosses with the tutorials and examples
that they can run in theirown laptops before to move
to the real or thefinal production system.
And Intel has this possibility also.
Developers can accessIntel Developer Cloud
to test multiple hardwarebefore to buy that hardware.
(11:53):
That is really cool, andalso accessing accelerators
and accessing also thelatest, for example, AI PC.
So we are provisioning alot of tools to developers,
and also we have, I almost miss that,
but we have an amazing repository
where developers can testthe latest AI trends.
(12:15):
So we have OpenVINO notebook's repository,
where if something happenedtoday, literally in two days,
we will see the notebook withthat specific model for sure.
This is for the open-source community.
So you can test there,for example, Llama 3.1,
YOLOv10, and the latest AI trends.
(12:36):
And this is a great tool.
And the most important thing is
we are not forcing developersto buy specific hardware
to run those models as well.
So developers can also run these models
in the actual hardware that they have.
We are also supporting ARM,
and we are also supportinga variety of Intel hardware.
(12:58):
Also integrated GPUs,
that is the most usage,
an integrated GPU, thatwe can see in the world.
- Yeah, it's great thatyou are making it easy
for developers to getstarted with the equipment
or hardware that they have.
And a lot of the kits and the challenges
we were just talkingabout, and repositories,
these are ongoing thingsthat are available
(13:20):
to developers at any time.
But I'm thinking about, I know recently,
which probably feels likeforever ago, you were at CVPR,
and there was a competitionand challenging going there.
So that's more of aone-off timely challenge
that is available sometimesto developers going
to these events, havingthese things happen.
(13:41):
So I'm curious, Jason, 'causeI know the company was also
at that event, but there'sother events that you guys host
or that you're at that havethese developer challenges.
I'm curious, what wouldyou say is the importance
of developers going to these events,
engaging in these communities,
and participating in someof those competitions?
- Yeah, it's a good point, right?
(14:02):
I mean, like even just before CVPR,
Voxel had our first in-person hackathon,
actually in New York City,
and it's that type ofengagement where we really see
like excited developersengaging with new technology
and then really trying towork together on new teams
to kind of solve a new problem.
That was really fun, butI think like one key angle
(14:24):
for developer events isobviously education, right?
Like learning new things.
And I think if you take my earlier answer
about how AI has evolved
and think about like a keytrend for the future, right?
Like a key trend that we're seeing
for the future is languagecombined with vision
combined with new compute capabilities
and openly available dataand these foundational models
(14:47):
to really tackle new problems
in what at Voxel we call visual AI.
I think we're going to seeincreasing contributions
to that effect, but how do you do it?
What do you do, right?
One has to go to developer events
or other types of conferenceslike CVPR or whatever, truly,
to really stay abreast ofwhat's happening there.
I mean, for me it's, in some sense,
(15:07):
the educational sideis very natural, right?
I'm a faculty member. I teach.
I'm not teaching right now this year,
but last year I taughtintro to computer vision.
So like three hours a week I was doing
this developer event, in some sense,
for 300 students to learnabout computer vision.
So I think one thingwe've learned at Voxel
is this AI space is evolving so rapidly
(15:28):
that it seems like everyone,
even faculty members who'vebeen in the field for ages,
we're in like constantinformation-gathering mode.
It's impossible to stayup to date with everything
from like cutting-edgeresearch papers, on one hand,
all the way to what are the new APIs
and libraries that you have tothen, that you have to learn.
And so to do this, at least at Voxel,
(15:48):
what we've tried to do ismaintain like a weekly,
at least one per week,if not more per week,
sort of technical output thatin some form of an event,
like different formats,
that really allows thecommunity to stay engaged.
So we have an eventscalendar at voxelv51.com
that we can include in the show notes.
(16:08):
I think we have something liketwo dozen events scheduled
between now and the end of the year.
Just personally, for example,every Monday at noon Eastern,
I maintain these open office hours
where kind of anyone can sign into them.
They're on Zoom.
We talk everything from like...
A couple weeks ago we werereviewing someone's paper,
(16:29):
and we went through slides
and like actual like technicalmodel all the way to like...
Oftentimes I get asked like,"This is my first time thinking
about getting into computer vision.
What should I look at first?"
Right? So pretty broad.
But we have some hackathons,virtual meetups, and so on.
So I think that it's like raw education
just about foundational capabilities,
(16:51):
but also these developerevents really help engagement
just from like a staying upto date with what's happening.
- That's great, and that's awesome
that you have those open hours
that developers can justjoin and start to learn.
I'm curious because obviouslythere are virtual conferences,
then there can be conferencesin different parts
(17:12):
of the world, and it canbe tough for developers.
They can't go to all of them
or there's just so many outthere, it's hard to choose from.
Is there anything coming upthat you want to call out
that developers shouldhave on their radar?
Or is there anything,
any other resourcesavailable to them online
(17:33):
that you think that theyshould take advantage of?
- What Paula was sayingearlier, being open-source
is like the gateway tofostering innovation, right?
Like our software atVoxel51 is called FiftyOne.
It's on GitHub.
We have the permissive licensing
for the open-source component of it,
which is basically like oneuser, one machine local data.
(17:56):
You can fork it. You can submit PRs.
We make releases.
I think it's on the orderof every one to two months.
Every release that wehave has some content
from our community,and we've been educated
so much over the last fouryears since we released it
about from community needsand community contributions.
Most recently we have this newfunctionality called Panels,
(18:20):
which FiftyOne is basicallylike a visual component
as well as a softwareSDK for doing the work
that we're talking about here,
like data and model refinement,
but with Panels youcan build functionality
for the front end withoutknowing how to write React
or JavaScript or anything with UX.
You can write it right inPython, and all of a sudden,
you can still enhancethe GUI functionality.
(18:42):
So I think those are great ways,
actual events but alsojust becoming a part
of open-source projects is another way
to really to get involved inthe developer ecosystem for AI.
- Yeah, absolutely, andI think it also helps,
companies like yourself whohave these open-source models.
You might not have picked up on something
that somebody in the developercommunity picks up on,
(19:03):
and they can really bea part of that community
and make changes and pointthings out and contribute
to companies and projects like yourself.
So it's always great to bea part of those discussions,
see what's going on,hearing what developers
are talking about as well assome of the ongoing challenges
(19:23):
that they're facing in these spaces.
Paula, I know OpenVINO,
there's a huge GitHubcommunity around there,
and you mentioned a little bit of the kits
and some other things thatIntel offers, but I'm curious,
in what other ways doesIntel foster that innovation
and that communityengagement for developers?
(19:44):
- That is a great question
because we have been workingso hard on that part as well.
So we have multiple ways.
So we are creating multiple ways
to create this innovation with developers.
So we have one program, thatis the Innovator Program,
where we have multipledevelopers around the globe,
that they can try, theycan test technology.
(20:05):
They can make their own applications
and they can share that with us.
So just stay tuned, for example,
in my LinkedIn or in my network as well.
So we are highlightingsome of these innovators.
This is one thing that we have.
And basically they createtheir own repository.
They fork their repository andthey create new applications
or improve the applicationwith the contribution.
(20:29):
So another thing that we haveis Google Summer of Code.
We have a program with Google every year
where we have multiple proposals
and we have several developers
around the globe as wellworking with us for three months
with different mentorsin the OpenVINO team.
And, for example, youmentioned about CVPR.
(20:51):
So we worked with Anomalib.
There is a library that also we have
in the OpenVINO ecosystem.
And we have two proposalslast year about Anomalib,
and one of these proposals,
the student that was involvedin Google Summer of Code,
and the mentors and the professor as well,
they create a paper.
(21:13):
The paper was submitted in the workshop
of the anomaly detection,
Visual Anomaly InspectionWorkshop at CVPR,
and that was accepted.
So we are closing also the gap
in between industry and theacademia with conferences.
So we are also participatingwith the students,
developers in those conferencesthrough programs as,
(21:35):
for example, Google Summer ofCode, but no more than that.
For sure, we are moving so fast also
in relations with universities,
what kind of things we canwork with universities,
helping them to create some research
and research proposals thatIntel also can support.
(21:55):
At CVPR, we are sponsoring as well
the challenge in this workshopabout anomaly detection.
We try also to in invite developers,
and we create a marketing campaign
around the challenge to invite developers
to participate in that challenge.
We receive more than 400 participants
(22:19):
and more than 100 submissions.
That was an amazing and remarkable number
around maybe one month anda half that we received,
and we can see how the knowledge is moving
in using anomaly detection.
For sure, talking about OpenVINO,we have multiple things.
As I mentioned before, OpenVINOis an open-source tool,
(22:42):
and we have a repository
when we have differentkind of contributions
depending of the product.
So we have OpenVINO, OpenVINO notebooks.
We have OpenVINO build and deploy.
In that repository,OpenVINO build and deploy,
You can find all the Edge Reference Kits
that we have been talking about today.
(23:04):
OpenVINO notebooks, youcan find the tutorial,
and in the OpenVINO repository,you can find the API.
So we have a huge ecosystem
where we are trying to touchnot just the inference part,
also the training part withanomaly detection, Anomalib,
and also OpenVINO Training Extension.
So we have a huge ecosystem
(23:25):
that I really want toinvite all the developers
and all the people thatare watching this podcast
or listening this to this podcast
to visit those repositories,
visit the organization,"openvinotoolkit" in GitHub,
and you can find all the repositories
that I'm talking about.
- Absolutely.
It's exciting hearing all ofthese different resources,
all these different waysdevelopers can get started.
(23:48):
So I'm excited to see movingforward what types of solutions
and innovations developerscontinue to build,
and I hope they take you guysup on some of these events
and meet you, whether that'sin person or virtually.
I know sometimes it can be intimidating
when you're gettingstarted in these areas,
but having companies like Voxel
and Intel support developers,that's great to see.
(24:11):
And I also saw, Jason,
in addition to the virtual office hours,
there's availability todo one-on-one meetings.
So if developers feel intimidated somehow
or don't want to ask aquestion in a group setting,
it's great that you guys aremaking yourself available
to help developers whenand where they need it.
(24:31):
So I want to thank you both again
for joining us on this podcast.
Before we go, if there'sany final thoughts
or key takeaways you wantto leave developers with
as they go on this journey,engage with each other,
and engage with yourselves.
Jason, I'll start with you.
- Great. Yeah, thanks very much.
So I mean, first parting thought would be
(24:52):
that I think I just wantto express my thanks
to the developercommunity that we've built
over the last four or five years, right?
We wouldn't be where we aretoday without the community.
It's such a vibrant and rich environment.
But like second thing isthat actually we're hiring.
So we're hiring developers.
(25:12):
I mean, actually across theboard we are as we grow,
after we closed our SeriesB earlier this year,
but for this conversation,machine learning engineer roles,
both for core engineering work
as well as developerrelations work, right?
We believe in developers so much,
so we hire individualsthat are fully trained
and can write papers, canwrite code, and so on,
(25:33):
but their role isactually building bridges
with the community.
And then maybe just the lastparting remark is that we...
As a company, we are open-source driven,
but we do actually havedozens of customers
that use our commercial enterprise version
that we call FiftyOne Teams.
It kind of like relaxes thatindividual user local data work
(25:55):
and allows you to develop thesame functionality together
in teams, in the cloud or on-prem.
And we love to engage inconversations around FiftyOne Teams
as well with with your community.
We have customers, many ofwhich are in the Fortune 500,
but across manufacturing,security, automotive, right?
(26:15):
Like pretty broad basecustomer base. So thanks.
- Yeah, absolutely love tohear about job openings.
It shows this space is growing,
this space is becoming important,
and some of the innovationsand transformations
that we talk about on "insight.tech"
wouldn't be possible without developers.
So exciting opportunityfor anybody listening
(26:38):
to go join the Voxel51 team.
Paula, always love havingyou on the podcast.
Thank you, again.
I feel like every conversation,
there's something new to talk about,
something new happening in the AI space,
so curious what our nextconversation will be about.
But before we go, arethere any final thoughts
or key takeaways youwant to leave with us?
- Yes, for sure. Sofirst of all, thank you.
(26:59):
Thank you, Christina,for create this space
to talk about what we have,
and thank you also to Voxel51.
We have been creating agreat relation with Voxel51.
Different conferences, we tryto share some space together.
And this also talks pretty well
about that we have the real intention
(27:21):
to work in the open-source community.
So we are open to work with all of you,
try to find the best path to developers,
because here the mostimportant thing are developers.
So the company for sureis really important.
We have a lot of thingsto learn from the company,
what kind of products we can provide,
what kind of tools we canprovide to developers.
(27:43):
And always we are thinking that, I mean,
we need to enable you to usethis hardware in software
that we can provideand you can accelerate,
you can improve yourpipelines and your workloads.
That is the main intention.
We have right now a lot ofthings to share with you.
So we talk about OpenVINO,Edge Reference Kits,
(28:05):
but more things are coming in the future.
For example, we have the new,
you have the new AI PC that you can try.
We have a new enginein the microprocessor,
that is the NPU, Neural Processing Unit,
that we can also expediteand accelerate part
of the conventional and generative AI,
conventional AI, generative AI,
(28:27):
process with that small device.
This is one of the thingsthat we can talk about
(chuckling) in the future,Christina, for sure.
Thank you again,
and I'm looking forward toconnect with all of you.
- Absolutely, and you talked about earlier
how some of these innovations
or these tools you haveavailable is making it easy
for developers to start working
(28:49):
no matter what hardware they're using,
and the AI PC just makesit that much easier
for the AI development,deployment, performance
of your solutions, all that great stuff.
So I know Intel has a lotof resources around AI PCs
that we'll make sure toprovide to developers as well.
But thank you both againfor joining us today.
Thank you to Intel and Voxel51for these great resources
(29:11):
and communities you'vecreated for developers
and spaces for them to getstarted and get that support.
Until next time, this hasbeen "insight.tech Talk."
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.