Latest / Agentic Conversations (formally mlops.community) / Vision and Strategies for Attracting & Driving AI Talents in High Growth // Panel // AIQCON

Transcript

Vision and Strategies for Attracting & Driving AI Talents in High Growth // Panel // AIQCON · Agentic Conversations (formally mlops.community)

0:09
Welcome, everyone, and thank you so much for joining this panel.
0:13
We have some amazing people that we're going to talk to.
0:16
So yeah, would love to kick it off.
0:18
Ashley, why don't we go ahead with some introductions?
0:21
OK, sure. Is this on?
0:22
Yes. OK.
0:23
Hey everyone. I'm Ashley Antonides.
0:25
Right now I'm the AIML Research Director at 26 Technologies.
0:30
For most of my career I've LED and a mixed research data
0:35
science, ML developer teams leaning a bit more toward the
0:41
R&D side. I spent about 10 years in the
0:42
DoD and a research organisation, have been involved with a couple
0:47
of early stage startups that were very R&D focused.
0:50
And then now the work that we do is is very our research focus.
0:52
So we do a lot of work with DARPA and other government
0:56
research sponsors, and I will pass it to older.
0:59
OK. Hi everyone.
1:00
I'm Olga Brigavia. I'm a Vice President of AI at
1:03
Smartlink. We are AI powered translation
1:07
platform. My in tech career has been a
1:09
natural language processing from day on one early days of machine
1:13
translation, summarization, sentiment analysis and so forth.
1:16
And slowly you're working my way into more diverse AI.
1:20
Awesome. And my name is Shelby Waklu.
1:22
I run my company called Shelby Ventures, working as a
1:26
fractional executive with early stage startups.
1:29
Formerly, I was the head of data at a company called Strava, and
1:33
now I'm also a best selling author on a book called Self
1:35
Advocacy. So with that, let's kick off our
1:39
topic for today because we have a lot of accomplished people
1:43
here with us. And I think AI in general is a
1:46
very hot topic. Everybody's focused on it,
1:48
everybody's talking about it. But the thing that people tend
1:52
to focus most on is the technology aspect.
1:55
And what I'm really curious to learn from both of you is what
1:58
about the people aspect? How do you find the, the, the
2:03
exciting talent in this field? How do you sort of set them up
2:07
to build the solutions that we want to build?
2:10
And Olga, I'd love to start with.
2:11
You, yes, it's already hold the microphone again as well.
2:14
Yeah. I mean, people could not be more
2:15
important. And we all know, I mean, well,
2:17
from the same space here, we'll know how hard it is to find
2:20
talent in our space, how hard it is to retain talent.
2:23
So maybe I'll go into how do you keep people?
2:26
I, I think it's all about an interesting job in this modern
2:29
time and day. It is so exciting to be in the
2:31
area, in the field that actually just keeping people
2:34
intellectually stimulated. I mean, all the other perks
2:36
aside, I think basically, basically I think that's key.
2:39
And I know we'll go into more detail later, yeah.
2:44
So I think in in addition to thinking about sort of technical
2:47
skills when hiring to, I'm sure a lot of us think about just
2:51
overall team composition, how the team is gonna work together.
2:54
I think that a learning culture is really important, especially
2:57
now there's so much new coming out every day.
3:01
So how to build that team culture where everybody is like
3:03
reading papers and OK, ask questions and really
3:07
collaborating and building off of each other's energy.
3:10
I think we'll talk about kind of diversity later, but I think
3:12
that's also an important component.
3:14
Just thinking about the AI systems that we're building are
3:17
interacting with people and how people can bring different
3:21
aspects to thinking about how what we're building is going to
3:26
affect them and and different people across different walks of
3:29
life. Whatever we build, it affects
3:31
people. Companies always try to have
3:34
that perfect alignment with, you know, company goals, product,
3:39
road map, and that talent that is internal.
3:42
So how do you sort of think about, you know, that org
3:44
structure that helps you support that perfect alignment or that
3:48
perfect teams divided up into those functions that actually
3:51
support those goals? So one thing that comes to mind
3:57
is, is actually just areas where I have seen maybe friction
4:00
between between teams or as we're trying to align and say
4:04
toward the business goal of delivering a product.
4:08
I think maybe one kind kind of common division is between like
4:11
the developer team and the data science team.
4:13
Maybe that will change over time as there's more blending between
4:16
the roles, but it can, I've seen a lot of teams that are
4:19
structured as like the dev team versus the ML team.
4:22
And so I mean, I think 11 way to address that is if you have sort
4:26
of an overall leadership role, whether it's the chief product
4:28
officer or the CTO who really is coming has empathy and
4:32
background in both, both perspectives and is helping to
4:35
build that infrastructure between the teams for model
4:39
delivery and the, the different sort of approaches and workflows
4:43
that those those different roles have.
4:46
One other area where I have kind of seen some friction is as more
4:50
and more companies or organisations are realising that
4:53
it's important to have a chief AI ethics officer to be
4:57
addressing ESG, having maybe more data governance.
5:02
Sometimes a technical team can see that role as almost like an
5:04
adversary or like an auditor who's coming in.
5:07
So what I've seen worked well is where you can bring that person
5:09
actually be part of the technical team and there's
5:12
really a dialogue about what are the true requirements from that
5:14
perspective so that the technical team understands that
5:18
and then that is incorporated into the overall kind of
5:20
deliberate delivery life cycle. OK.
5:24
I completely agree in a couple of things I wanted to add.
5:27
We all see that at the end of the day, it's not that difficult
5:30
to hit an LLM API and deploy, deploy an application
5:34
internally. And what I see also, and I don't
5:36
know whether it's friction or opportunity, you often see,
5:39
especially in large enterprises, this accessibility that can
5:43
drive inequal parts, collaboration and friction.
5:45
So what I'm seeing a lot is in the product team, they just
5:47
like, I mean, I'm in the translation space.
5:49
Hey, let's just plug in ChatGPT and let's translate everything
5:51
with GPT. At the same time, you discover
5:53
that in digital marketing, there is a separate initiative because
5:56
everything is so accessible. So I think it's super essential
6:00
to have again, same thing centralised governing, governing
6:04
and while again, fostering creativity, making sure that
6:08
there is internal alignment as opposed to billions scattered
6:12
individual initiatives. I also think what's essential
6:15
now, again, there are so many aspects of equally developing
6:18
and deploying AI, You really need to have a champion in every
6:21
function, dev, product, marketing, compliance, HR, every
6:26
function now plays a role. So again, it's all about
6:28
identifying the needs and the role of every function and
6:31
making sure that there is alignment.
6:34
I I love that idea. Like have a data or AI liaison
6:38
in every department. And actually to add to this, I
6:40
was on a panel with a lady from IKEA and they indeed have the
6:44
digital ethicist that actually looks, and I'm pretty sure it's
6:47
not unique. She was speaking about
6:49
specifically about this role, AI digital ethicist that's really
6:53
looking after ethical deployment and distribution of roles within
6:56
the company. That's really cool.
6:58
I also think like, you know, often when we talk about ethics,
7:01
one of the topic that comes up is that if we have very
7:04
homogeneous teams, we end up having products that have a
7:08
little bit of a blind spot on what is possible.
7:11
And the common answer that a lot of people say, you know, can
7:14
help solve for that is to have more diversity in the talent
7:18
profiles that you have in AI. What about that aspect?
7:21
How do you prioritise it in your teams?
7:23
What do you think about it? And you know, most importantly,
7:26
what happens if we don't prioritise it?
7:28
What happens if we don't prioritise having diversity in
7:31
AI? Maybe I'll go first.
7:33
I mean, we've all, I'm pretty sure we've all seen similar
7:36
research that if you don't ensure diversity equally on
7:40
model data collection, model development and model
7:43
implementation team, you actually do get this one sided
7:46
view. And we have empirical evidence,
7:48
especially when it comes to different kind of cultural and
7:52
yeah, all sorts of biases. Actually one of the techniques
7:55
known techniques of mitigating data biases and algorithmic
7:58
biases is actually ensure there's diversity on teams.
8:01
So where I work, for instance, what we see is we have very,
8:04
very diverse representation. Again, I come from the language
8:07
AI space, one of our chief data scientists.
8:10
She's an, not only is she a Chinese lady, she's also a young
8:14
mom who is raising a toddler. And it is pretty fascinating
8:17
things that she notices from just her language perspective.
8:20
She comes from a completely different language group, so she
8:23
catches hallucinations and thinks about things that I would
8:25
not even think about with my language background.
8:28
But on top of that, she's a young mom.
8:30
And what she says, which I absolutely love, is she says,
8:32
look, I do two things at home and at work.
8:36
I'm talking to a toddler. So my prompt engineering
8:39
actually is very identical to what what I'm telling my
8:41
toddler. Give very specific instructions
8:43
to somebody who doesn't quite understand but wants to learn.
8:51
So just thinking about, so acknowledging, I think, I think
8:54
diversity on, on AI teams is really important.
8:58
How to, how to build that It's, I mean, there's so many
9:00
different aspects of diversity, but how, how to build that can
9:03
sometimes be hard. You can right Often like you
9:08
look at teams and they may be homogeneous.
9:12
So I, I personally feel pretty strongly that the hiring manager
9:16
plays a really important role. And one thing I like to ask when
9:19
we're interviewing is just some basic question about kind of
9:22
your question, like what, how do you think about diversity?
9:25
How do you think it's going to be?
9:26
How will it play in your role? And just even that simple
9:28
question can actually you can you can tell if the person has
9:32
even like thought about this this at all.
9:35
But I have seen where if you have, you know, someone who
9:39
maybe hasn't thought about that their their team as they grow
9:41
it, especially if you're in like a high growth area where
9:43
somebody's going to be hiring a lot.
9:45
You can kind of see how that composition is a certain way
9:47
over time versus someone who is really intentional and really
9:49
wants to to have that be an objective as they're building,
9:54
as they're building their team. Yeah, I'll just leave that for
9:57
now. Yeah, I think, I think another
9:59
aspect there is also sort of, you know, we, we don't talk
10:03
about educational diversity a whole lot, like a lot of job
10:06
descriptions in AI and data in general, they still sort of
10:10
require certain types of education.
10:12
They're not as open to people who've come from sort of a boot
10:15
camp or self self learned, self learned background.
10:19
And I really think like, you know, if you've excelled in a
10:22
different field and then you choose to come to AI, like
10:25
that's a huge, huge benefit for AI that, you know, you get to
10:29
learn from an expert in a different field who's also now
10:32
applying, applying to that that area.
10:36
So, yeah, so I think. Just I think that's a really
10:38
good point too. I think, yeah, the role of also
10:42
just the domain expertise and just the lived expertise of the
10:45
folks on your team and how to leverage that I think is really
10:48
important. I really like what you said
10:52
about the outside perspective, fresh perspective, right,
10:54
Because a lot of people have been in the same field for years
10:57
and then somebody comes in from different industry somewhat
10:59
adjust and skill set. They will have all the
11:01
capabilities of learning all the on the job, right?
11:04
But they will bring the perspective that we would not
11:05
have. Absolutely, Absolutely.
11:08
So sort of shifting gears from, you know, we've all had a tough
11:13
time hiring, hiring and identifying the right talent,
11:17
but for the talent itself. So there are so many people who
11:20
are, you know, excellent AI innovators, and everybody's
11:24
pitching to them. Everybody is pitching to them,
11:26
like, come work with me, my problems are great or my company
11:29
is great. So what are your strategies from
11:32
separating yourself from the noise?
11:33
How do you make sure to set a really clear vision and a clear
11:38
pitch so that AI experts choose you over the next company?
11:43
Maybe I'll start here. I think you make the message of
11:46
impact. What we're doing is real and
11:49
it's going to be impactful and it's going to drive outcomes.
11:53
So I think it's much easier to attract the talent if you say,
11:55
hey, this is where we're going. This is the impact that we're
11:58
going to have on our customers society.
12:01
And I mean, you can take it to any degree and any scale.
12:04
And if they see that they're working towards a specific
12:06
tangible goal that contributes, I think at least that's been
12:10
something that's been working for me.
12:11
It's got to be interesting and you know where you're going with
12:14
this. So that's one of the techniques.
12:15
And obviously, again, if you're inviting somebody into a diverse
12:18
team with great culture, that's something usually what I would
12:22
do, I would do around or what we would do around of interviews.
12:24
So management interviews, senior management peer.
12:27
So the person really gets to meet everybody and be motivated.
12:32
I think so for a lot of the teams that I work with, again,
12:34
are very, are very research focused.
12:36
So there is a lot of interest and motivation or questions
12:39
around what conferences am I going to get to attend?
12:42
Is there an opportunity to publish?
12:43
Is there an opportunity to at least do maybe like blog posts?
12:45
So having that record of like, hey, our team is published here,
12:48
you will have an opportunity that's prioritised.
12:51
And also just people talk about what the training budget is or
12:54
what the training availability is.
12:56
So folks know like, hey, I'm going to be able to go and keep
12:58
my skills up and, and go to conferences is really important.
13:02
And then I like to talk to to the learning culture to about
13:06
kind of what cadence we have for internal reading groups, for
13:09
workshops, for cross training across maybe other groups within
13:13
the organisation. I think that's really key, just
13:15
that folks know there's an opportunity for them to keep
13:19
upscaling and keep learning new things.
13:21
OK. And actually, I'm very curious,
13:23
but in your team, do you like just suggest that people in the
13:27
team go for conferences or do you almost require it?
13:30
Like is, is industry thought leadership, you know, part of
13:33
their criteria for career laddering?
13:37
We don't require it, but we actually as a team, we try to
13:40
identify maybe three or four kind of key conferences
13:43
throughout the year. I find that there's value if you
13:46
can send maybe like a few folks to the conference, especially if
13:49
there's concurrent tracks. They have an opportunity to come
13:51
back and kind of talk with each other.
13:54
So we try to do that kind of collectively of like what are
13:56
gonna be like the CVPR Nurips that we're gonna send the team
13:59
to, OK through the course of the year.
14:01
Yeah, yeah. Maybe I could add to this and
14:03
there was a person in the room way more qualified to speak on
14:05
the subject cuz Jen there manages all of our conference
14:08
submissions. But no, in reality, I think
14:11
we're almost crossing over to please publish.
14:14
So people know that they will be given the budget to travel to
14:17
conferences, but also know that they will be encouraged, greatly
14:20
encouraged to publish. One thing is because we are in
14:23
practical application of AI, we actually have something that,
14:27
and those of you from academia would probably agree, we have
14:30
the data, we have the field data that we can share.
14:33
So quite often when we go even more so to research conferences,
14:36
we bring something that quite often folks from academia might
14:40
not have as much exposure to. So we're usually a very welcome
14:43
addition because we come with practical, real life case
14:45
studies. So yeah, encouraged and welcome
14:48
to publish. So yeah.
14:50
Awesome, awesome. That sounds lovely.
14:52
All right, so the next topic, I know Olga, this is a favourite
14:56
of yours and I know Ashley, you're passionate about it too.
14:58
But we know that AI implementation still requires a
15:02
lot of contributions from the human in the loop.
15:05
So how do you how do you sort of think about that?
15:08
Like, you know, is, is there specific strategies that you
15:10
have for again, identifying the right people like with with the
15:14
right us? Yeah, we, I mean this, this
15:17
conference is about quality in AI and we still agree that
15:20
despite all the mitigation techniques models do hallucinate
15:24
data is biassed. So you do need an even more so
15:27
again, in our industry, in language industry, you do need
15:30
human in the loop and you'll probably need human in the loop
15:32
in the foreseeable future. Equally for direct assessment,
15:35
for human ranking, for validation, for post editing,
15:39
for fact checking the not an issue.
15:43
But one of the key factors there is quite often that would be a
15:46
smaller gig. Quite often that would be
15:50
sometimes it can feel like a repetitive job, like if you need
15:53
to click X number of images and decide whether it's, I mean,
15:57
right now we're beyond a cat or a dog, but there are other
15:59
criteria or if you need to validate an output.
16:02
So it's very important to keep people motivated, those human in
16:05
the loops to actually be there and contribute.
16:08
And I mean, first of all, I mean, it's a combination of it's
16:10
fairly easy money to be made on the side.
16:13
But also I think what we're learning is essential to give
16:16
the right tooling. If you give the right tooling to
16:19
those people who perform the human in the loop tasks, if it's
16:22
easy to rank prompts, if it's easy to assess the output, if
16:27
it's easy work surface, that's usually what keeps people
16:29
motivated because they feel that they are productive, they feel
16:32
that they're doing something interesting, and they feel that
16:35
they're given the right tooling to do their job.
16:39
I guess for us, depending on the project, we usually go through
16:42
kind of three different options for data curation.
16:46
One being, do we hire somebody into the into the company?
16:49
That's usually not the course that we go because when it's
16:51
sort of questions around what is the growth path for this person,
16:53
maybe it's something where they come in as a data creator or
16:57
labeler and then there's an opportunity to grow into more of
16:59
an ML engineer role. We've talked about, not in my
17:03
current company, but I've been at companies before where we
17:06
actually have almost like an internal crowdsourcing system,
17:09
which I which maybe isn't totally scalable, but it is kind
17:11
of. I do like when the ML team has
17:14
an opportunity to label, just to actually go through some of
17:18
those edge cases to put together the guidelines and then think
17:21
about, OK, is this now at a point where we need to scale
17:23
this up and pass it out to an external labelings labelling
17:28
company or partner. And I know there are challenges
17:31
with that too. There was a great panel this
17:32
morning on data quality and labelling and all those
17:36
challenges. But those are typically like the
17:37
three courses of action that we are kind of courses that we look
17:40
at. Just on the subject of tooling,
17:43
I wanted to comment. I don't know if we have any
17:45
Kalina folks here. I'm a huge fan of the product
17:48
and I swear I'm not like I'm not affiliated with a company in any
17:51
form of shape, but when it comes to model evaluation, I I think
17:54
that's highly recommended. OK.
17:56
So yeah, whatever. Yeah.
17:59
And I think I think, you know, back to the earlier point like
18:02
exposure, the more exposure that people on your team have at
18:05
conferences to other tools or to new newer technologies that
18:09
people are using. Like it's, it's really great
18:11
when people come back to the company and say, hey, there's
18:13
all this other stuff that we have out there that can not just
18:16
make a life better, but it'll be more accurate.
18:19
It's better for our product, it's better for our, you know,
18:22
for our customers. So it's it's great to be able to
18:25
identify, identify those pieces. Yeah, I guess one thing that's
18:31
on everybody's mind, whoever came to this session, where do
18:34
you actually source talent from? Well, I mean, we might be saying
18:40
the same thing. I mean, so I think we're
18:43
laughing because it's it is a challenge.
18:47
So I think we'll see. There's always the sort of first
18:51
tier of, you know, who's in, who's in my network or the
18:54
network of the, the people like within the group or within the
18:57
company. I think that second to that is
19:01
like we, we develop a lot of partnerships with universities,
19:04
go to conferences, that sort of thing.
19:05
So kind of building overtime that broader network of partners
19:10
who may also have students or folks who are who want to come
19:13
in. And then sort of scaling from
19:16
that too, like we have worked with I mean both with our sort
19:21
of HR departments and also with maybe recruiters or headhunters
19:24
too, which is maybe a little bit more hit or miss too.
19:28
But that is one way of of scaling the approach.
19:32
And I think there is just been for me most effective working
19:36
with them to very clearly articulate what the requirements
19:39
are because maybe they may not understand or kind of know like
19:41
what the use case is. So it takes maybe some iteration
19:45
in that case, but yeah, I just want to acknowledge like it can
19:49
be challenging, especially going back to sort of like diversity
19:53
too. So how do you how do you recruit
19:54
for that and and find those pockets as well?
19:58
Yeah, I agree and also. I think being present at
20:02
conferences indeed matters because if you, if your
20:05
presentation, your company presentation is super
20:07
interesting and you're delivering some, if you're
20:09
sharing some interesting insights and you're sharing
20:11
whatever the latest be it's research or be the product
20:14
feature. Usually those in the room who
20:16
are looking for the next most interesting thing would react
20:19
and quite often we would be approached actually like, hey,
20:22
what you presented really resonated with me.
20:24
I really want to, I really want to be a part of this project.
20:27
Do you guys have anything interesting happening?
20:28
Same usually applies actually to LinkedIn activity.
20:32
If somebody of the company's posting interesting content and
20:35
the company leadership is posting interesting content,
20:38
again you get followers. Eventually those followers will
20:40
graduate or will be looking for the next opportunity and they
20:44
will come. So quite a bit of inbound if you
20:46
do enough outbound. I think that's what works and
20:48
internships always work magic. If somebody comes and works but
20:52
looks for an internship, actually we just hired couple of
20:55
interns as FT ES as full time employees after the internship.
20:58
So but it is hard. I'm not going to say that, you
21:01
know, you wake up in the morning and hey, here is your senior
21:03
data scientist with natural language processing background.
21:06
No, not quite, but they're always.
21:09
Yeah. You know, so my, my last
21:11
company, like it was a consumer tech tech company, very popular.
21:15
So I, I think that was the one time in my career that it felt
21:18
very easy to hire because like, if people love the product, they
21:21
will come to you. And like I, I worked in other,
21:24
like most of the other companies I worked at were B to B.
21:26
And, you know, nobody wakes up thinking like, oh, I'm going to,
21:29
I'm going to see if they have a job for me.
21:31
So it's I, I like, I like the idea of like, you know, putting
21:34
yourself out there. Maybe it's projects that, you
21:38
know, even your consumers can work on or even somebody else
21:41
can collaborate with. Like those can become spaces
21:43
where you identify the talent, then it comes to you.
21:47
So once you find, you know, once you find that talent, this is an
21:51
area that is facing very rapid changes.
21:54
Like, there's a lot of stuff changing everyday.
21:57
How do you encourage your team to stay sharp on their skills?
22:00
How do you sort of, you know, keep rejigging the team as
22:03
needed with the changing environment?
22:05
And how do you actually grow that talent?
22:08
Maybe I'll start here. I really agree with Ashley what
22:10
you said about learning culture. I think it's super important to
22:13
do internal webinars, internal knowledge sharing.
22:16
Like you know, I don't know if you guys are familiar.
22:18
There's archive.org, which is like a big source of NLP
22:22
research. Like you would just divide
22:24
articles and just say like, hey, such and such person will report
22:27
on this and such and such person will report on this and let's
22:30
brainstorm. And then always, since we're all
22:32
in the field of dynamic research, I think just making
22:35
sure that people follow you don't follow the next shiny
22:38
object, but you definitely want to be appraised of the
22:40
state-of-the-art. And I think that again, if
22:43
somebody's research topic is pretty cutting edge, the rest of
22:46
the team will learn from them. So I think, again, it's almost
22:50
impossible. I think many of us probably have
22:52
been in the situation in the previous conference I was
22:54
speaking at, the submission deadline was a week, a month
22:58
before the conference. I showed up with a flash stick 5
23:01
minutes before my presentation because everything I was
23:03
producing was becoming obsolete every morning.
23:06
So I think it's, I mean, that's just the fact of the matter.
23:09
So you don't chase, but I think you put just the tracks of these
23:13
are the topics we're going to be, we're going to be
23:15
researching and making sure that somebody really goes in depth
23:18
and can coach the rest of the team.
23:22
I mean, so yeah, I definitely agree on kind of internal brown
23:25
bags workshops, having internal reading group.
23:30
One thing that has worked or I found has worked pretty well
23:33
across a couple different teams is we call it like a micro
23:37
sabbatical where folks can say here, here's like a topic that I
23:42
really want to learn about. And I think it's going to be
23:43
most effective for me to take a week or two to go research that
23:48
and maybe maybe the end product is a white paper or a prototype.
23:52
Very least they're going to present back to the group and
23:54
share what they learned. Sometimes that's more effective
23:57
than trying to find a training class, especially for some of
24:00
these newer topics. And it also gives them like the
24:02
actual hands on experience. So that's something that I
24:06
found, yeah, to be pretty effective across our teams and.
24:08
Maybe I could add to this. And also I think it's very
24:10
important that the company is willing to invest and invest.
24:14
I really mean dollars. I really mean dollars in terms
24:17
of allocating certain budget for people to take courses, for
24:20
people to get certification, to be able to expense their books
24:23
if they are, if they decide to buy books.
24:25
And I think also people stay motivated when they see that the
24:28
company is really investing time and effort and actual money into
24:32
their growth. I think that also helps.
24:35
Yeah. And to your earlier point, like
24:36
not just invest this for the data team, but like maybe across
24:40
the company just to improve the education and culture related to
24:44
the topic so that, you know, even if it's like, hey, product
24:47
managers, how do you engage with this topic?
24:49
Designers, how do you engage with this topic?
24:52
And nobody cancelled the good old hackathons, so there is
24:54
always that. Yeah, so I, I think you know,
24:58
you know, you brought up the example of like as you submit
25:03
new stuff, like it goes out of fashion or something like that.
25:06
So I think that's one of the components of like how do you
25:08
actually keep your team motivated like this?
25:11
This landscape changes very fast.
25:14
Companies are in almost sort of an AI arms war of sorts where,
25:18
you know, maybe you're working on something cool in your
25:20
company and another another company gets to it first.
25:22
Like, how do you actually keep talent motivated and excited
25:26
about what they're building and the solutions that they're
25:29
bringing forth outside? Well, again, I think I mean,
25:34
from my perspective, it's again, it's all about the outcomes.
25:37
Like if the person delivered, developed a prototype and
25:40
suddenly your automated scores are through the roof.
25:44
I mean, that's very inspiring, right?
25:45
When you did something and you really see, hey, I really did
25:48
something awesome and you really see that the metrics, like you
25:51
met all the metrics. That's one place where like
25:53
people really see the outcomes of, of what they do.
25:57
And then I would also say like, I mean, in my industry, it's a
26:00
lot of it is about quality, language quality.
26:02
It's super arbitrary. But when the person sees that
26:05
their or the team sees that their work really turn the dial
26:09
of the needle, turn whatever that was on on the customer side
26:13
and we see that our customer C sat goes up customer
26:17
satisfaction and the customer is able to meet their KP is that
26:20
also keeps the team motivated because they did something good
26:22
for the customer. It really has practical,
26:24
practical implications. I agree.
26:27
I think you mentioned earlier kind of the impact.
26:30
I mean, in addition, in addition to just the continuous learning,
26:34
I think really seeing what the end value is, it is really
26:39
motivating. I know for us again, like we're
26:41
coming from more of a research perspective.
26:44
So often that means getting to the end of a project and then
26:46
it's transitioned to that customer and really seeing what
26:50
the operational or the end uses for sure.
26:53
Awesome. So I think this this next topic,
26:56
maybe it resonates with a lot of people, but how do you how do
27:00
you keep your team focused on the current projects?
27:04
Because there's a lot of distractions, There's a lot of
27:06
shiny, shiny objects in the horizon that people may chase
27:09
after. How do you actually keep people
27:11
focused? So it is challenging, I guess
27:17
maybe in in some ways it's sort of redundant with what I've
27:19
already said is like having those outlets for for still like
27:22
exploring new papers, being able to do some prototyping, some
27:27
sabbatical, going to conferences so that there is space for that
27:31
and bringing and then also thinking about how to bring that
27:33
back to the current work as well.
27:35
So it's not so much, you know, someone's 100% on the current
27:38
project and can't spend any time thinking about newer things.
27:44
It's sort of a high level answer.
27:45
But yeah, what do you think? I actually agree, I think to a
27:48
great extent is I think it's Google, right, that has this one
27:51
day for your personal project thing.
27:54
I mean, you don't necessarily have the luxury to do it every
27:56
day, but you can like somebody on our team, he loves to write
27:59
like a long esoteric threat test on different topics of natural
28:03
language. Sure, do it if you want to do it
28:05
one day a week, right? Every dissertation you want,
28:07
we'll listen, we'll talk, but then let's just make sure that
28:10
the other four days of the week we're actually working towards
28:13
towards teams goals. But it's extremely difficult.
28:15
It's super hard to stay focused for US as leaders or for the
28:18
team when like there is a new model released every day or a
28:21
new method released every day. So what do you do with this?
28:23
People do get distracted. So I think it's really against
28:26
setting the KPIs. Let's meet the KPIs and let's
28:28
have fun outside of it. Keep the threshold for the KPIs
28:31
and then then you earn yourself some some time off for the shiny
28:35
projects. All right, so just to close this
28:38
out, you know, there's a lot of competition for niche skills.
28:42
How What are your go to strategies for retaining AI
28:46
talent? I'm good.
28:49
OK, I think I'll just repeat myself.
28:53
I think it's really just I mean, first of all, there are I mean,
28:55
there are we're all humans, right?
28:57
We live in the society. So there are all the perks, the
28:59
compensation, right, and all the all the things that are just
29:03
native to you. You have a job, you have a job,
29:05
but I really think you retain the talent by keeping
29:09
researchers and product managers and developers lives Interesting
29:14
if there is enough to satisfy their intellectual actual
29:17
curiosity. I think people stay that and
29:19
culture. We all know that people don't
29:21
leave companies. People leave their bosses quite
29:23
often, right, and leave their teams.
29:26
So it's essential to keep that there.
29:28
They cannot be any us and them. There cannot be any competition,
29:31
but that's on us as leaders to make sure that that's fostered.
29:35
I agree. And I think at this point it's
29:37
reiterating what we've already already discussed.
29:41
I'm trying to think in terms of, yeah, really, really building
29:44
that culture where everybody's supporting each other.
29:47
I think just for folks to be able to see that they can
29:50
continue to grow throughout their their career.
29:53
One thing I guess we haven't talked about is also if folks do
29:56
want to take some time to go get a degree or something like that,
30:00
having support for that in the company, whether they need to
30:02
take time off to get a full time degree or do something that's
30:06
also part time and they're, they're also working at the same
30:08
time. So I I think it's, yeah, just
30:11
identifying those opportunities for continuous growth.
30:13
As they say, culture eats strategy for breakfast.
30:16
Yes, Awesome. Well, thank you both so much.
30:20
This was wonderful and I appreciate you both sharing your
30:23
thoughts and thank you to the audience for just sitting with
30:26
us. Thank you.
30:26
Thank you.