let me oh my god did it go computer froze whatever like lately my computer just been frozen I think we're on. Are we on? No, it's still processing. I'm so sorry. Normally it takes a lot faster. But today, I don't know what's going on. Maybe it's... You were seeing life on your end? oh my god i'm so sorry i'm like i do not know if we're live wait let me just refresh oh oh we're live we're live okay perfect hi sabrina welcome to
Hi, Stephanie. Welcome to Winter with Grace. I'm so sorry. I'm like, I just want to make sure it's like actually live and I'm like having a major heart attack right now. But anyway, we're live. In the past like 45 seconds. Okay, great. Okay, so I think you have a really interesting career. And, you know, you were at Google for over 19 years. And then before that, you know, you have been a software engineer or system engineer for various different companies.
want to start with like, what were some core lessons that you've learned early on in your career that you kind of took to later on running tech organizations? I think early on. I think early on, I realized that you really have to drive your career, you know? I think... Right out of school, I thought, oh, you know, your manager is going to guide you. And, you know, sometimes that happens. But sometimes they.
want to point you in a direction that maybe you're not interested in and so i think you have to stay very focused on what it is you want and not let other people make those decisions for you you know people will steer you and it's all very helpful and i've had great uh leaders and mentors over the years but very early on i realized that sometimes people would steer you in something that wasn't going to work for me. And so I had to find my voice and really understand what motivated me.
Because there's a lot of things you can do, but you really need to make choices on things that will keep you motivated. Because our careers are 40 years. Even when you're... super lucky. It's probably going to be 40 years. You know, you can count your school years if you want to cut that down, you know, like count it how you want. But our careers are long and you want to.
be motivated to to stick with it yeah totally were you always loving engineering or like um i guess like for your current job since you're the cto of gitlab and i guess like What are something that you feel like are different from, I mean, obviously it's very different from like writing code day to day, but I guess like, what does a day in your life look like as a CTO of GitLab?
put a day in the life of a CTO at GitLab. I think the great thing, and I don't know that I understood this before I got the title, but. I'm now accountable. One, I do love engineering, but I'm actually in engineering because I'm a problem solver. And I think this field... has a wealth of problems to solve. But the day in the life for me is I have to stay very focused on the business strategy.
you know the whole company what are we doing and while i lead the engineering division which is all of development operations and customer support As CTO, you're actually responsible for the overall business strategy. So I need to understand what's happening and go to market. And what are they hearing? What do they need? I need to understand. everything happening in legal, right in finance, in product, as well as engineering. So like, I love the fact that I have to have the big picture.
top of mind every single day um and then for my day job like i'm looking at like what is the customer experience overall Not just in the day, but like, what am I hearing? What is our survey feedback? I meet with customers and talk to them about what's going on for them. I have to... you know obviously development execution right what are the big things we're doing are we on track for our milestones and then you know production health because i have all of operations so what's happening there
um and then of course my people right not my people but like the org overall you know how's the health what's the mood how are people feeling because you know there can be a lot of pressure And as a leader, it's super important for you to understand how, you know, how is the pressure today? Right. And what can you do to mitigate that or reassure people or just check in on them?
And then, of course, there's the cross-org health, right? Like, how are our relationships with all the other divisions here at GitLab? So I love it. It's a lot. It's a lot. I didn't realize how much it was going to be, honestly, because I had run a much bigger org before than I have here. But and, you know, I had like business strategy and stuff like that, but I wasn't the one responsible. And so it can be really intense. incredibly rewarding on those really really good days grace
For sure. I thought like it was so interesting. Like, okay, maybe we could start with like from a product level, like what is GitLab? And then we could talk about like the business model since you mentioned about like, you know, you were taking care of a lot of like these like business strides. strategies and, you know, running software engineering operation and customer support teams. And why don't we start with like, what is GitLab? And then we talk about the business model.
Great. Yeah, thanks. GitLab is a DevSecOps platform. And what I like to say like we're software engineers who build software to help software engineers build software. And GitLab is the most comprehensive software development lifecycle platform. And that's one of the reasons I decided to come here, because the power of a platform to really help teams be more efficient, right, in how they work and to get their job done and to have an impact.
on their business was very compelling to me. And so we start very early on in the software development cycle. So like planning.
right how do you do your planning how do you break down the things you want to build we break it down we put it into epics and issues and that way all the different teams can collaborate together on what they're going to build and then of course we do source code management so as you are building your code right you're saving it your code reviewing with with your your peers and your teams
And then you move into like testing your code, documenting your code, you know, checking for security vulnerabilities. We do all these scannings. We have, you know, our runners, our runners go out and can.
do all the all the automation that you know software engineers need to be done validate that the code is is there all the integration tests all your smoke tests and then of course you can push it into production and then you know we can monitor it and give you analytics about how all of that is going so yeah it's a really powerful platform for sure i wonder so like how i guess like for a person who i think a lot of people use
good lab obviously but like i guess like how does a business side of things works i guess like who would you say are your target like icp and like how do you think about from like a business perspective as a cto You mean the business perspective of GitLab's business or how we serve our customers' business? Both. Maybe we could start with GitLab's business model.
Yeah. So one, we're an open core platform, which I also liked a lot. And I can talk about what that means. But, you know, we're working with customers who have to deliver.
their software safely. You know, we try to learn like what problems are they dealing with now what what are the drivers for their business what are they concerned about about delivering software you know and think about like where their developers are how automated is their system today you know where their needs are and we work closely with them we sell getlab obviously to the customers and it's tiered right based on the needs of the customers i think
A big topic today is cybersecurity, right? And, you know, I think... early on in in in GitLab really people were just looking for source code management and then maybe the deployment but then over time there's lots of different personas that are involved with building software and that's when we started to shift left like more into the design phase but today it's really about the security scanning and catching security problems
early on um and so depending where companies are and what their big focus is are they just trying to keep track you know source code management or are they trying to do the full life cycle um and then you know we we sell to them we actually have three different platforms that we offer so we have a sas
offering, but you don't have to, not all companies want to move their software development into the cloud. And so we offer GitLab for them to self-manage if they choose to do that. And then we have a fully managed. service where they are on an isolated instance and we manage it for them. We manage their GitLab for them. And so all three platforms.
Totally. I wonder since like, you know, you're running engineering operation and customer support, like how do you structure your day? So like, for example, on the day to day engineering. So obviously a lot of people producing code and then.
On the operational side, what does it kind of look like for you? Obviously, customer support is also a really big part to make sure the clients are adding on to things that they... um want to add or like knowing how to use and everything but i guess like from a day-to-day perspective as a cto how do you kind of like manage these three big bracket how forward
well i'm super lucky that i have an amazing leadership team and i do think that that's super important right um you look i need to know what's going on but you know i'm only one person um and so i you know in my first year here i really spent time you know working with the leaders to make sure i understood what their needs were they understood what my needs are and
because we need to communicate a lot. I would say I am on Slack all day. You know, I didn't use Slack when I was at Google. And when I first started using Slack, I was like, wow, this is like a major DOS attack. Right. I am in so many channels. I've figured it out. I've got my priority channels in there. And so it's really working with the leadership team, my VPs, my directors and the managers as needed. Right. But it's important that, you know.
We understand how to all work together. I never want to undermine my leaders. I'm just there to support them. And so, yeah, it'll depend. Like some days I'm like spending more time in the AI.
Sometimes more time with the infrastructure team, right? Like, how are we scaling? Managing three platforms is... is challenging and there's a reason a lot of companies are deciding that like everything just needs to move into the cloud and it's because it can be very complex i actually was pretty impressed when i came to get lab to sort of see how We actually managed the code base. We have a single code base. These are not three different.
codebases, that's very powerful. It actually also requires the teams to be incredibly responsible for how things are working. together and then we're also open core right so the community can contribute back to us and we have to treat those you know in the same way that you know the same code reviews the same diligence the same testing and all of that um and you know that takes a lot of time
from teams and diligence. And I'm just, I feel so fortunate to work with so many people that have such a deep, deep sense of responsibility. to making sure that this is scalable and that it works well and that we're being really responsible because the impact.
affects the customers um i also think open core helps us a lot right there's a lot of people looking at our code commenting on our issues gitlab is incredibly transparent so if we have an issue open like the community can see it they can comment on it they can tell us whether or not they think we're triaging something appropriately But that makes us better. I strongly believe that. We were very transparent within Google, but not external. And I think it's been a transition for me to come here.
be so transparent but i love it because it it forces you to be better For sure. I wonder, so like, you know, you have been at Google for over 19 years. And why don't we talk about like, you know, I guess, like, how do you if let's say you have like three. We separated the 19 years into like three segments. Let's call it like six to seven years each. And then like, what would you say were the core learning journeys like look like for you? Maybe like, let's say the first.
six years you just entered as an engineer and then in the middle you transitioned into leadership in the end you were like the full-on vp of engineering of the core infrastructure i guess like what does it mean for you while you were you know working on um well i was a manager before i joined google and i actually joined google to not to get out of management actually I, you know, I always and even I think when I was a manager, it was very important to me to be.
a very technical manager to always have my own projects and i did at google for a very very long time um i think that was great about google that that you know they wanted you to be technically engaged. I think it makes you a better leader to do that. So I think early on was about, and you have to remember the company was so much smaller.
Like when I joined, we would have, you know, the developers had test machines. We had test machines under our desks. Right. And you would realize that maybe something in production was actually. running from someone's desk like that's actually how small we were when i first started um we fixed that very quickly just to be clear um and then you know i was mostly focused because at that time when i joined it wasn't what it is today. Android was a recent acquisition when I started Chrome.
was an idea you know gmail was not what we used internally so it was very early on um and i think everybody was in it right we wanted to be a different kind of company and everybody was trying to figure out like what does that mean and we were very empowered and we had a lot of autonomy so being an engineer at google was amazing you know it was i feel really
lucky to have been there at that time. And then later on, what I realized was I was actually a good leader. And I thought that, you know, googlers you know that's how we refer to ourselves googlers deserved good good leadership and and i knew that i could i was really good at getting my work done and i knew that i was good to get
help other people get their work done and i think as we were growing because i was a man i wasn't a manager for probably four or five years um while i was there i can't remember now um and so i went into management because i thought i could
increase my impact and do a good job leading a very empowered autonomous developers to challenge maybe even what they thought they could do um and then so line manager for a while i think a big breakthrough probably i had been there but a year five or six i moved to gmail um and you know, during like just this crazy growth period and we were driving so much changes and I learned how to, you know.
i think what we would say like you're changing the engines when the airplanes in the air because millions of people were on the platform at the time i was in sre um i worked very closely with the developers. Gmail was probably my favorite experience because of how collaboratively we took on really, really hard problems. well together, both developers and SREs. And we had to foster that trust relationship. And I was a big part of that.
and i think it it sustained for a long time and i also think because we pushed the boundaries i think a lot of the work we did there influenced a lot of the products at google and i think that's why over time my My portfolio expanded to, you know, I got to work with all the product areas and I feel grateful like every single one of them had. had something that that they taught me pretty fun.
that i lean on today and i think about today and when people ask me like how do i scale our systems i have so many examples of like how do you change the back end
you know, without taking downtime, right? Like how do you responsibly migrate user data? These are things that, you know, were really, really fundamentally important that we all out great responsibility to our users and um and i i keep that to this day you know and then i think moving into it was grace i have to tell you people remind me all the time that
you know as i was growing as a leader people were like oh do you want to be a director at google and i was like oh never be a director at google you know i just never thought that was possible the the leadership at google was so amazing that i
i didn't self-identify in in that group and then uh and then i became a director um and you know i i did not say i won't be a vp but i felt like i wasn't going to be a vp and then you move into the vp and i think each of those different executive levels you take on this new level of responsibility that's more than it's your org right like when you're a director you have to be making decisions that affect the business and then when you become a vp
you have to realize that your decisions affect the company right and so it's this this larger sense of responsibility and You need to be very focused to make sure that every day you're thinking about the company. And I think it's important to understand you don't not think about the people. You don't, you know, like when you're a team manager, you're thinking about your team and the work your team is doing.
and then you have multiple teams and you're thinking about the people the work that they're doing and then the bigger picture that bigger umbrella And then director, you kind of think about the business and then VP, you think about the company. But you still have to think about all the other things too.
For sure. I think at one point like you were managing a 3,500 people team at Google and how do you think about like I guess like how do you separate these like different departments and then I guess like. What would you say are some management lessons that you've learned throughout managing thousands of people? I don't think I ever got to 3,500. I don't know. I did not get to 3,500. I think it was 2,000.
My org was about 2000 at the largest. I think that when I had 2000 people, that was actually all of Google's consumer products. I think it was more than 90% of Google's revenue came out of teams. I think there's a couple of things, you know, some people think the bigger your org.
the better. And I actually think the opposite. I think you can do so much with so much more with with leaner teams just because so much coordination between people it can be really really challenging and it takes a lot of overhead And it takes away from your ability to do technical work. So I think the important thing is to make sure your teams are.
are as small as they can be without overworking people. I kind of had a strict rule or guideline, not strict. I tried to make it strict, but you know, you got to be flexible, is I looked at how people spent their time.
and i would want 50 of their time to really be spent on the long term right and then 30 would be you know midterm and then i wanted 20 to be short-term decisions right those are things that like you're reacting to and you're you you know they come up you get escalations or something you didn't expect happens and in order to kind of keep those ratios you have to really start to map out your work
and so i was a big advocate for roadmaps you know i like teams to have roadmaps and then i like groups of teams to have roadmaps and then larger roadmaps and and as a leader you know you can be very prescriptive but my style is much more so you know set the guard rails for people and give them the information that they need but
As an engineer, I wanted to make decisions about what, why, how I was working. And so I try to do that for my team too. And I think that that's... as a leader super important because There's some really smart people and I can make a lot of really I can make the decisions for my org. But if I'm the only one making it, then if something happens to me, it fails the organization.
So I am constantly thinking about like, how do I scale? How do I increase the impact? And then how do I make sure that what the org that I'm building is sustainable without me, which is important for leaders to think about. I think that's really interesting. I like what you said about like having the work system without you. And then I guess like just like self.
or like automate the entire process. And speaking of that, I would love for you to chat about, you know, what are some like AI trends that excites you? And then what are some AI tools that you're building, either internally or just like a lot? leveraging outside of, you know, the company? I guess, like, how do you think about what are the AI tools that will help you as a CTO to function from day to day basis? Yeah. So what am I excited about?
look i think ai is is is uh evolving so quickly it's amazing to me i mean i'm pretty fortunate because you know ai or ml was you know common in google right like we used it before and and i think that uh i think people didn't believe it would be as um accepted you know by consumers or by everybody in the way that it has been. So I think that's been really, really surprising.
And I love to see what people are doing and using it for. The models are just getting, you know, really amazing and they're always surprising. um and i used to enjoy at google when you know you'd ask it to summarize something and i love to find all the mistakes right like i just i thought oh look this is It's putting things together incorrectly, but in really funny, funny ways.
So something, but what I'm really excited about today, obviously the developer tools, we have all the tools to help people write code, generate code, chat, ask questions, and that's all. good, you know, it helps the individual. But what I'm really excited about today is agentic AI, you know, and we're building that here. It's called Duo Workflow.
The things that we're able to do, because I always wanted to be more efficient in my job. And then as a leader, I wanted to make my teams more efficient. um because i thought it freed up their time right to on those long-term, right? Like I was always aiming for at least 50% of the time thinking about the future, really building for the future. And I think agentic AI actually. enables that kind of ratio work.
right because in order to get rid of these interrupts or things that just pop up you can have the agents watching for those things and taking care of those things so you're not surprised i'm so excited about you know a team here we were going to do an accessibility fix it and i think you know anyone who's had you know a user facing application especially something that goes into like schools or certain industries
you need to make sure that your code is very accessible and meets all of the different compliance requirements. But over time, I mean, almost every product I've ever worked on has had to have some kind of accessibility fix it because people forgot or it wasn't included explicitly in a spec. And so you go through your backlog, you identify all these problems and, you know, I challenged the workflow team, you know, like, you know, could, could we use workflow?
to do the accessibility and they did and i was just like this is a game changer right that the ai can go and identify problems and you know give you suggestions to fix it Because what that means is you could put this agent on all front end changes and have it find these things proactively so you're not leaving them in a backlog for a fix it where everybody surges to try and fix it.
And so I see agents doing these kind of things that pop up and annoy us or agents watching for security alerts that come in. give you like hey here's all the things that we need to fix or and by the way here's all the suggested fixes that you just need code review test or whatever. And it can test them. It can do all these things for you before you push it out to production.
Because, Grace, I mean, think about things like Log4j and how many people had to stop everything and go figure out how to address this large. Broadly impacting security alert. And we've shown that we can actually do that in our platform. And that's just really, really exciting to me. Right. I think when you free up. you know, engineers times, it really gives them to space, space to think about what those next breakthroughs are going to be.
So I don't believe all the hype that the software engineer is going to go away. Right. I think we're going to free up their time to do things that we don't actually know we need yet. Totally. I wonder. OK, so.
thinking about in the future of like a developer um what do you think they would um what is like the new version of coding will look like so now it's like everybody write their own code and then uh check the code and then you know ship the code and then monitor everything and i wonder in the future like do you imagine like the co-pilot just write its own code or like um i guess like or for let's say git lab dual like do you guys like um envision this thing as like a uh you know i guess like you
helping you to check the code or would you say it's like, oh, everything will be auto-generated in the future? I don't know that everything will be auto-generated. You know what I think for sure in the next five years? i think people will be able to prototype things very quickly right i think that you know things that take a long time to design right like i was talking to someone last night actually and they were like
You know, I don't know what the product manager or UX person is going to do because right now they spend a lot of time in Figma. But now you can tell the AI like, hey, I want this, put a button here. And it just. built it all right so like imagine if you just get to talk to things see what you want and then say yes like it's just gonna speed up the time to innovate and while i think that ai can generate code for you and build things that you want i think that anyone who's worked at scale right
knows that like it's not just you know code completion right like there's so many big architecture decisions that have to be made you know like what does the storage look like and what's the apis for that storage look like in order to you know you're not Most people are not just doing thousands of users, right? Like you got to be able to hit millions and billions because that's how many people are online. And you have to be able to handle like.
For every user, they have a phone and a laptop and a tablet, you know, their TV, their hub, whatever they have, their watch. there's a ton of different connections and you have to be able to orchestrate all of that stuff so it's very difficult for me to imagine that all of those kind of decisions can be made by ai and the reason is grace like i've worked on a lot of different products and how you do this is different
right it's different on how the user interacts with it it's different with like where in the world they are and in order to deliver service that's available and also maintainable and sustainable and resilient. It's oversimplification to think that all of the applications and opportunities in the world just require you to write code because that is not what's involved with delivering software today.
um for sure i wonder so like one of the audience asked a great question so chris been asked curious what um what a team of dev look like when you have uh two thousand under management and then is a team of teams or each composed of like eight to ten people or like something else entirely um i think like another question i have
like piggyback on this is like would you say in the future of like in your future engineering team instead of like 2000 people would you say it's like maybe half AI tools and then half like human or like There's always rumors saying that the more junior stage of engineer are going to be replaced by coding tools. What does it look like for you nowadays?
Maybe from his question, you know, is there like a lot of micro teams versus, you know, big teams or like in the future, how does like the engineering team is going to kind of made up? You know, I think when you have, let's start with the first one. So when you have 2,000 people, look, that's not going to be a flat organization, right? be and you don't have in in my case i didn't have 2 000 people working on the same things right this is 2 000 people that covered
lots of different businesses and lots of different products individually. But let's say just 2,000 people, they're definitely usually in teams. You have management in there. Depending on how cross-functional it is, you probably have program managers in there. So you have your ICs.
They're all different levels. You have your tech leads, you have your managers, you have your program managers. There can be some strategy people, you know, like once you get over a thousand people, it's super helpful, especially as a leader.
kind of have a team around you because you know you're not just managing work um you have your people partners right that's like the hr people who are helping you you need they need to help with like the learning and development and the performance management and so you kind of have a little bit of overhead i try to keep my overhead like about 10 to me that that's
that's the sweet spot um especially if the managers are are technical which which i want them to be so sometimes um if you have a single product area you don't want to just have like oh every team is six people right because that doesn't allow a lot of growth and that's a lot of cross-team coordination which is really hard grace so but you don't want to have everybody try to do everything that breaks down too and so that's how you figure out like
how do you how do you break up you know a thousand people into teams it's not a size game it's not strictly numbers right it's about what problems are you solving What how can they work with autonomy? Where are the dependencies? And that helps you lay out like, where do you need senior managers? Because if you have teams that have dependencies on each other, it's super helpful to have a single leader that can.
coordinate and negotiate right when there's conflicts and there's always conflicts when you have people involved. Now going to the future of AI, you know, I do think that you'll have these agents.
that will be working um we have seen that we can train the agents in the same way that we onboard individuals but i i think that people should not think ai is equivalent to a developer right because the ais can do much more um so i would say if you have 2 000 people and you're really leveraging ai what that means is that's probably
maybe five years ago, it could have been maybe 4,000 people, but you probably never would have done 4,000 people. So your 2,000 people will probably do a lot more in the future. than the 2000 people from five years ago. Right. But think about this. When when Here's how it works, Grace. So let's say I want to do something new. I will go to my team and say, hey, let's do this new thing. 100% of the time, maybe it's 90, but it's very high.
They will say, I need headcount. Right. I need new people. And so that's how orgs grow. Right. It's because the amount of work grows. So I think in the future with AI. The teams will grow slower, but they're still going to grow because we're going to be creating more. And when you're creating more, like there's complexities with that of how do you deliver that? And, and even.
I think you should even go more broader in engineering because the ecosystem around a software engineer, I think that's important to think about. So like I said, you have these orgs and you have ICs and you have program managers.
when you're delivering products you have product managers right and then you have go-to-market people and then you have sale you know you have the field team and customers customer support and all of these things are you know could technically be done by one great person, but break it up because it's too much of a cognitive load. and and things you need to know for it to be one person so i think what happens is the entire ecosystem is affected when you bring ai in right like product managers
can do their jobs quicker because the AI can shortcut it for them. Right. And the same with program managers, like you might need less program managers because the AI is going to coordinate a bunch of stuff for you. But in reality, that 2000 person can just deliver more with the help of AI and because they're just going to be more efficient and you're going to have more. headspace to really think about like what is the
next thing that would change the world, right? Or whatever your mission is, right? And companies have lots and lots of different missions. You know, some are working on climate and it kind of opens.
so many avenues like I felt people ask me all the time why I stayed at Google for as long as I did and I did it because there were Google was investing in lots of different things so i wasn't just ever on gmail i got to work with the youtube team and i got to work in ads and i got to work in search and google invested in like what is sustainability right and how do we
How could we be more responsible stewards to the environment? And I was super motivated by that. And so all of these big problems came up. And I think you could say that was special about Google. But I think in the future, all companies are going to have the bandwidth to just actually be able to do this. And it will be good for their business. And it will be good for their people.
For sure. I wonder, what is it meant to run the core infrastructure at Google? So, like, I know everything is built on top of infrastructure, but, like, I guess, like... yeah like i'm just out of curiosity on like what does it kind of like intel to like build and build the entire infrastructure and then since google has been there for a long time as you mentioned like well you joined 19 years ago or
i guess you worked there for 19 years so like that's a long uh that's a long time and i wonder uh i'm sure the infrastructure kind of shift towards like from the beginning you know there was like five code or two now it's like you know um 100k colder I don't know so how do you think about like I guess like how does the infrastructure even change and then how do you manage the core infra like what does that intel to yeah um
So I didn't run all of it. You think that I could. But, you know, when you're running infrastructure. So just to. explain a little bit about what core infrastructure is at google it's the very product oriented infrastructure like google has underlying infrastructure like the networking and the storage but the core infrastructure would be you know the authentication systems the abuse systems analytics you know really data oriented um oriented systems and so infrastructure for me
is you're solving lots of teams problems at the same time, right? That a lot of the things that are in the core infrastructure at Google were things that were previously very specific. to a product area um you know like data analytics started with in ads right because we needed to do that but they created something so powerful that rather than everybody have their own analytic systems we should we should share that
And how you do logging, like having a lot of people log data in different things, that's a risk to your company. And Google understood that very early on and understood the power of what.
where the data the data and the logs were so that became something centralized and then all the tooling could be built on that and then all of the value of that data could be available to other people and you could control that data access, you know, centrally to make sure that, you know, user data or the different kinds of data were not put at risk, right? with a very open culture like Google has. And so I think when you're running infrastructure, you have to be a very good problem solver.
Right. You have to be able to understand the requirements of different users and find the commonality and the opportunity. And then you also have to be influential. Right. You know, developers are incredibly opinionated and passionate about their beliefs and they want it the way that they want it. And I think when you run infrastructure. Some people can get away with just mandating like, no, there's only one way to do it. But the most powerful infrastructure really.
which Google has a lot of it, really understands how to navigate all the different requirements, find the commonality and then raise the bar for what is possible. And and so I think the people who are really good at infrastructure are big, big picture thinkers and actually. are very analytical so they can really understand how to navigate and see the commonalities between all the different inputs that they will have. Let's say like, I mean, since like, I mean, I read.
somewhere that like you know gitlab reported total revenue increased like 31 since last year and then or like year over year and then i wonder when it comes to like increasing revenue like as a cto how do you think about building like from like adding a different product or like different revenue like like how do you think about like what are we
gonna build and then since like ai is like dominating right now and then how do you think about like you know we just talk about like using ai to like improve code but i guess like how do you think about building the ai product that actually will be mattering to the developer community. I think that it's so important. You can come up with some really great ideas in the abstract, Grace, but I think... In my experience, it's really thinking about the user.
um and like what their experience is the good thing for us is is we are the user right like i said we're the software we're software engineers too but that's not enough Right. And so I think in order to think about the value and and, you know, making revenue on it, you have to make sure that you're meeting the needs of the customer. I think it's very special. GitLab that we're able to meet the customer where they are. And we're also are very invested. We do something called co-create.
Right. And that's one we're open core. So if you if we had a gap and you wanted to contribute, you could. But sometimes what we do is like we actually see the benefit that of a unique. need for one customer. And so we spend time really understanding that and then we'll co-create things together. So we did a partnership with Amazon. we genuinely came together and thought this would be good for both of our customers and, you know, and built, you know, an integration with Duo and Q, right?
um because we saw the value of the customers to be able to use both both both of our companies technology um and so when you know you can sell something, but in order to see the growth and the stickiness with that customer, you need to be very invested in like, what is their return? What is their return? Not just what it's yours.
and you know as a cto that's the most important thing for me right it's a real balance and and it can be a really good idea but i'm not motivated tech for tech's sake like that's not who i am But, you know, there are those people out there and and it's amazing. I actually try to balance both. Right. Like what can we sell? But also like what if they're going to buy it.
what's the return for them, right? I have to be like, does this make them more efficient, right? Or does this make their resource costs go down? You know, what is it?
a value to them so we can make sure one that we price it correctly and then the other side of growth that the cto influences is actually the cost the operating cost for delivering your services and your software and so i work closely with my operations team and and you know genuinely all of my leadership to make sure they understand what is the cost of their decisions you know and and i had to do less of that at google you know adwords is just an amazing
money machine. But when you're in a smaller company, you genuinely need to understand what decisions you're making and what are the costs that you're bringing to your own company. Right. And then everything that I invest in, I want to deliver value to our customers and in return, we would get revenue from that. And so that's sort of how I think about that. For sure.
I wonder, so like, how do you think about the current, I guess, like, AI startup landscape? Like, what are some products that you're excited about using at GitLab? as a cto and then what are you excited about like just building in-house because there's a lot of like different ai solutions and um you know people are building and then just curious on like you know what to build and what to buy Um, you know, I, I, um, that's such an open-ended question because there's so much.
There's so much out there. Maybe we can start with like as a CTO, your tech stack, like what does your tech stack kind of look like to build stuff? And then like, or like what does GitHub? Oh, we use GitLab. We use our own tech sack. So, I mean, I interact with it. I think for Duo, the things that I do the most, because there's so much going on.
I am so thankful that like you know we have our issues and our issues give the whole history of like what's going on and then there's all these opinions in there and then there's decisions and it's all kind of embedded in different spaces and i just go to duo chat and say summarize issue for me and then it gives me like
all of the summary of the discussions that is so helpful right because i have to make decisions very very quickly now you know the summarization it you do normally have to look at the details to make sure you understand what the summary is i would say i love you know outside of get lab i would say uh i actually love zooms summarizing of notes and just so you know like at GitLab when I joined like everyone's writing notes right every meeting we pretty much would transcribe
the meeting because we work asynchronously so if people couldn't make you know a meeting they actually could read the notes and see everything that was discussed in detail right and i think today you know products like zoom they not only transcribe the meeting for you which allows the people who are in the meeting to participate much more because we weren't all taking notes with you know anymore um but i love that it summarizes that and i would say that that's been a delightful uh
outside of GitLab AI feature that I was like this, I love this, right? It saves my time and it allows our meetings to be much more engaging. Everyone is much more engaged. um what are something that like for example let's say you're building a new product or feature like
I guess like what kind of customer input would you, I mean, how do you value the customer input? Let's say if you travel with a customer and then they ask you to build a certain product or feature, like what are, how do you kind of like take feedback from that perspective? I think it's, that's kind of a complex question, but I think the most important thing is to understand that people will tell you what they want.
Right. And I think when you're on the receiving end of that, if you take it literally, sometimes you don't solve the problem that they're actually having. Right. You hear what they say, but you need to spend time understanding why they think they need what they need. And that's because it makes sure that you implement it the right way.
um it's super easy to miss details and then you deliver you think you deliver what they want but you missed an important detail and you delivered something that they absolutely didn't want and so when you're getting requirements it's super important to understand and how they got to the conclusion for what they're asking for.
one because it makes sure that you meet all of their requirements even the unspoken ones and then the second reason to do that is sometimes you can see how to make it better right because they're focused on one problem And because you have a different perspective, you can actually see a bigger opportunity if you really try and understand. you know the details or the background for why they're they're doing what they're doing or asking why they're asking what they are um
One of the audience said, like, how do you stay focused or reprioritize on the product in GitLab, especially in the face of competition and slash like reprioritizing the product roadmap? Well, I have a great... cpo that i get to work with um and look i think especially in ai and i think especially in the last year and i don't think it's unique to get lab i think the technology
kind of changed underneath you. I would say normally in my history, you know, you'd have this new technology and you would start building for it, but it was kind of sticky, right? So really building something. brand new from finish to end you know kind of at scale for sure takes about two years right the interesting thing about ai is nobody has two years right like in a year
It has come, you know, like what would have normally been two or three years, maybe even more. I would say the last year, I'm shocked at how good the models are. So much better than they were a year ago. So it's progressing very, very quickly. So I do think in today's age you need to be able to pivot very, very quickly. And I think going back to making sure you understand the problem that you're solving.
And then the second thing is like, what's the quality bar and how is that changing? Because I think that probably was less of an issue in the past or with previous technologies than it is with AI.
um because you can say oh here's the quality bar but because the technology progresses so quickly people people change faster than technology does in in some ways because if you see something better we want better right everybody wants better and so you can set a quality bar um for you know january and by june everyone's expectations are higher and so i think in today's age like you constantly have to know like what were the signals that
that you set that quality bar against and make sure you're challenging them i used to do this a lot i used to force my teams to to really reassess the assumptions that they had worked from originally because you know assumptions change now i used to do it on like you know every few years but now it's almost you know on on
i would say monthly but probably if we're going to get to the daily where you're constantly going to have to think about what are the assumptions that you're working from or we're working from last week and and do they still hold true because so many the expectations are changing so quickly because the technology is changing so quickly i wonder so okay so i just want to be mindful of the time i didn't realize we're about to
uh hit the time okay so uh i have a one minute viral for you uh what's your favorite book my favorite book ever ever oh my gosh i don't know
What? Or what does your content that look like? I used to like... i used to like a lot of stephen king i read a lot of stephen king oh maybe you know what i think it's hitchhiker's guide to the galaxy i probably referenced that book more than anything else i would say any single book i probably reference that all the time oh fish right like i love i love the whole series actually i've read them all and i
I think about them all the time. So yeah, I guess that would have to be my favorite of all time. Love that. Who would you invite to your dinner party? I would like Barack and Michelle Obama. Hi. I just love the way they think and look at problems and really kind of make it so easy and accessible. Like they can have really...
well thought out opinions on very, very complex issues. And they're brilliant and impressive. And yeah, I would love to have a dinner party with them. I'd be so intimidated though, Grace. I'd be so intimidated. uh who made the biggest impact in your career um gosh there's so many so many people who have you know giving me great advice you know giving me amazing opportunities so many people on my team that I'm thankful for so I think it's hard I would say though like
somebody i think about a lot is megan smith you know i i had the fortune to work with her at google and then later on um when she was the cto of the united states i saw her conference and she was standing there and she had this long line of young you know early career engineers waiting to meet her. And she literally, I watched her for probably 25 minutes and she was just so present.
for every single person that she met that i was so impressed with that and like i i regularly think like am i being present with the person that i'm with and it's very hard because you know i'm always thinking about a lot of things but i do think that that makes you a great leader and so i remember that day a lot But it's hard. I mean, I got to work with Ours and Ben Trainer and so many people on my team that that, you know, have inspired me and made me better.
both peers and people on my team. So I feel bad about naming anybody actually. Where can we find you outside of work? Outside of work, the Lego store. What is going on in the background? They're like the whole Lego collection. They are all the way around me. I'm in my entire house. I have Legos in every single room just because I can't. i i hate to take them apart i love them so much um so yeah how long does it take you to build something um i don't know uh i
it's almost a DOS attack really, because once I open the box, I want to finish it. So I'll stay up all night. And just to finish a build. Oh, finish daily in the background, just every day. It does change. It changes a lot because I'll have something new and I want to show it. Yeah. What's your current?
build up like i mean is there like something that are you building right now i do have the new t-rex i haven't opened it because it's a really big big set and as soon as I open it so I can't open it for until the weekend oh my god I love that that's that's amazing actually I feel like that's kind of like building code as well but in real life you have to ensemble it and you have to like engineer the structure of it and then making sure you follow the step by step
The kits are easy. They come with a book. It's when I get something in my head that I want to build on my own that that'll take me a little bit longer. Wow. I love that.
like this is like the most like advanced lego collection i've ever seen in my life but i'm i understand the concept of like yeah anyway okay i love that addiction health um get lab lego but anyway okay so thank you so much sabrina for joining us today and then it was like such an awesome conversation and then just like looking at you seeing like the female leader in tech it's like super
impressive and i just like i really love your energy yeah thank you there's a lot of us people think that there's not um but there's a lot of us out there amazing okay so let me quickly end stream