LAURENCE MORONEY: Hi, everybody, and welcome

to TensorFlow World.

In this video, I have the privilege

to chat with Kemal El Moujahid.

KEMAL EL MOUJAHID: Hi.

LAURENCE MORONEY: And you are a director here at Google.

You've been talking.

And the keynote today all about community

and some of the great things that

are happening in community.

And I know you're fresh off, like, a trip around the world

right now.

KEMAL EL MOUJAHID: Yes.

LAURENCE MORONEY: We had a road show going on.

Could you tell us a bit about it?

KEMAL EL MOUJAHID: Yes, I'm very lucky.

This is one of the best part of my job

is to actually-- it is work for me to go and go

to these amazing places and meet with the community.

I had the privilege to go, amongst other places,

to Morocco, Ghana, Shanghai, and just

meet with the local communities and hear about what they

were building using TensorFlow.

And it was just incredible to hear the feedback.

Both the amazing things that--

the amazing use cases that they were coming up with--

things that we would have never imagined,

which is one of the beauty of working

on a platform like TensorFlow-- but also hear the feedback

and see how we can help organize the growth

because the communities have been growing so much.

It's global.

And so, how can we best help our community thrive?

LAURENCE MORONEY: So it's like, we hit every continent

in doing this, right?

And it primarily a listening tour.

KEMAL EL MOUJAHID: Yeah, so we had 11 events

across five continents.

And we heard so much feedback, which

really helped us think about how can we best help the community.

And it came down to three different things.

We want to help the community connect and discover

the rest of the community and see

what other people have been building

and share what they have been building.

I think also, we really want to help the community learn.

There's all these new wave of new developers and learners

who are new to ML and can contribute and want

to make sure to help them learn about machine

learning and TensorFlow.

And then we really want to help them contribute.

The community really wants to give back,

and there's some awesome mechanism to do that.

And so, we've been working on enabling those three things.

LAURENCE MORONEY: So then, for folks

who are watching this video who would really love to connect,

what ways can they connect?

How can they get involved in the community?

KEMAL EL MOUJAHID: Right, so the community has grown a lot.

I think now, we have passed 46 million downloads.

LAURENCE MORONEY: Wow.

KEMAL EL MOUJAHID: --2,100 contributors, and the community

is truly global.

This is something that we're really proud of.

And so, this is the reason why we wanted to organize something

like TensorFlow World.

We wanted an event that was really about the community

so that the community can come connect with each other

and discover what they were building.

And this process has actually started to happen organically.

The TensorFlow user groups, which

started appearing seven months ago,

and now we have 50 of them--

I think the largest one is in Korea with 46,000 users.

LAURENCE MORONEY: Wow, really?

KEMAL EL MOUJAHID: We have 15 in China.

And so, you know, chances are that, you know,

if you're watching this video, you

might not have a TensorFlow user group near where you live.

But there are TensorFlow users.

So if you're interested in creating such a user group,

we really want to help.

So that's one clear way that you can get in the community.

LAURENCE MORONEY: So like, one of the things

I've personally noticed is that, like, I'm

quite visible in the community between YouTube and teaching

classes.

So lots of people have reached out to me on LinkedIn.

And I see all these, like, little communities forming

of people helping each other, doing study groups,

and all of that kind of thing.

And you had mentioned that organic growth.

And I find that's really magical.

KEMAL EL MOUJAHID: That's really amazing.

And that's one of the things that

is most exciting about building a community like TensorFlow is,

you can see people interested in certain types of use cases

that you wouldn't have thought about, connecting with people

who are ML specialists.

And when that connection happens,

those amazing use cases are born.

And that's really something that we want to encourage.

LAURENCE MORONEY: So as well as individuals,

we also hear that businesses want to get involved

and enterprises want to get involved and learn

how they can use ML.

And I know there's some efforts in that space, too.

KEMAL EL MOUJAHID: Yeah, so we've

heard from businesses that, oh, we have this problem.

And we heard that it might be a good problem

to solve by with machine learning.

But we don't necessarily know where to start.

And on the other hand, we hear from partners who say,

well, you know, we're really good at machine learning,

and we love TensorFlow.

And you know, how you can you connect us with enterprises

who have those needs?

And so, that's why we launched the Trusted Partner Pilot

Program to help connect enterprises who have a business

problem that could be resolved with ML and TensorFlow

with partners who know how to do that.

LAURENCE MORONEY: And it's only just launched,

but I hear there's some success already.

KEMAL EL MOUJAHID: Yes, it's already just launched.

If you go on tensorflow.org, you can learn more about that.

But it's been some really interesting use

cases in the insurance industry and global met tech industry.

And again, if you're a business, and you're

interested in getting involved into ML,

I think there's this staggering statistic

that AI was going to add $13 trillion

to the global economy in the next 10 years.

LAURENCE MORONEY: Right.

KEMAL EL MOUJAHID: And right now, enterprises

are just starting their transition to machine learning.

So now is a really good time to start investing.

And if you need help, then you can reach out

to those trusted partners.

Also, another side of that, which we're excited about

is, if you're a partner yourself and you want to get visibility

and you want to help enterprises with their machine

learning needs with TensorFlow, then

you should totally reach out.

LAURENCE MORONEY: Cool.

So then, the second thing you mentioned

was, like, for people to learn.

So there's the folks who would love

to be part of these communities, but they're not there yet.

They don't-- they're not up to speed,

and they don't have the knowledge yet.

But we want to make that as easy as possible for them to learn.

So what's going on in that space?

KEMAL EL MOUJAHID: Right, so we've been investing

a lot, as you very well know.

LAURENCE MORONEY: A loaded question.

KEMAL EL MOUJAHID: Exactly.

On education material and especially with deploying

AI to just help onboard this new generation of machine learning

practitioners and make sure that, you know,

regardless of what their prior experience

with machine learning, they get to join the community

and contribute.

So we're really excited to see this new course launching,

and thank you, Laurence, for helping with that.

And we've really been inspired to see hundreds of thousands

of people just completing these courses

and then starting to apply machine

learning to real-life problems.

LAURENCE MORONEY: There's this, to me, this interesting thrust

going on, where there's all of this inspiration

about these things that you can do on the one hand

and then us trying to show people

that they can do it, as well.

There's no imposter syndrome.

People can learn.

People can do this for themselves.

And then, when these two things come together,

communities are formed.

Communities grow.

And then the magical stuff you've mentioned happens.

KEMAL EL MOUJAHID: Exactly.

I think there's-- you can get involved regardless of what

your prior experience on machine learning is--

again, because those tools are becoming so accessible, so

easy to use.

And there's so much educational content available.

We just launched our learn ML web page

on the website, which organizes all these educational materials

for you.

And it's organized by level.

So you could start with beginner or advanced right away.

And then it's all organized.

You can see the videos and the books

and the tutorials, which lead you

to being able to apply machine learning

to solve your use case.

LAURENCE MORONEY: And the website you mentioned

is tensorflow.org.

KEMAL EL MOUJAHID: Tensorflow.org.

LAURENCE MORONEY: You should definitely go check it out.

KEMAL EL MOUJAHID: Yes, go check tensorflow.org.

LAURENCE MORONEY: Now, beyond then-- like,

being part of the community and learning,

then, there's taking the next step

and actually contributing back to the community.

And there's lots of great stuff going on there that you

spoke about in the keynote.

Could you give us a quick summary?

KEMAL EL MOUJAHID: Yes, so I would say the first thing to do

is to reach out to your local machine learning GD.

We have, I think, now 126 GDEs.

We really love our GDEs.

They're amazing.

And they've been organizing so much awesome things happening

in the community, one thing being the Doc sprints, which

is a new thing this year.

We've been hearing those amazing stories

of groups organizing to translate documents or update

documents for 2.0.

They're going through power outages

and staying there and just making sure

that they finish their Doc sprint,

actually translating documents on their phone.

LAURENCE MORONEY: Wow.

KEMAL EL MOUJAHID: So I mean, this is incredibly inspiring.

And we really want to support that.

So if-- something that I personally heard when I was

in Morocco--

like, where's the documentation in Arabic for TensorFlow?

Well, you know, if you're interested in translating

the documentation, whether it's Arabic

or another language that's not available,

then you can reach out.

And the GDEs can help you organize a Doc sprint.

LAURENCE MORONEY: All right.

Part of what's great about this is

that with TensorFlow generally being open source,

there's so many ways that someone can contribute,

right-- doc sprints that you've mentioned.

People can check-- they can check source

code into the product itself.

KEMAL EL MOUJAHID: Exactly.

LAURENCE MORONEY: There is, like,

being able to build models and distribute them

via things like TensorFlow Hub.

There's just so many ways, right?

KEMAL EL MOUJAHID: There's just so many ways.

And that's the beauty of open source.

And to facilitate contributing code,