Cultivating Developer Joy with Rajeev Rajan, CTO at Atlassian

(Intro Narrator at 00:00:00) Today, we're talking to Rajeev, CTO at Atlassian, about how he's helping to create joy in the workplace. You're listening to Joel Beasley, Modern CTO.

(Joel Beasley at 00:00:15) You know, we get caught up in dumb things. We need to make sure our team is productive and joyful. And how do we do that?

(Rajeev at 00:00:23) Yeah, so that's a great question because when I came into this role a couple of years ago, one of the big goals we had was developer productivity. And obviously, everyone talks about we want developers to be more productive, and we did a bunch of things. We have done a bunch of programs to make developers more productive. But as we thought about it, we actually leaned into what we call developer joy, because, you know, productivity and efficiency, these are all goals. They sound very business-y, but as engineers, as developers, we want joy in what we do because that's what we do. We write code. We want joy. And what does developer joy mean? And to us, as we thought about it, it meant, you know, when you're writing code and you're in the flow of writing code, you don't want to get blocked.

(Rajeev at 00:00:51) The thing a developer hates the most is you don't know what to do next because you want to use a component, you don't understand that API, you have to wait for some engineer to wake up in a different time zone, and you're writing code at 3 a.m. in the morning, and there's no one to talk to. That stops your flow, that stops your creativity. And so we have done a bunch of things to make it so that you can actually find that component, you can find that API, and unblock yourself and keep going. And we actually have a product. Atlassian sells a product called Compass. Compass is like a place where you can register every component, every microservice in your company. And so anytime you start, go to Compass, find the component, find the API, find the documentation, and boom, you can get going and you can self-serve yourself and get unblocked. And so we've done a bunch of things like that as just one example where we want to make it so that the flow of creation, the flow of writing code isn't hampered. And that's how we have leaned into the joy part of it.

(Joel Beasley at 00:02:06) And so you took over for Sri?

(Rajeev at 00:02:09) Yeah, that's right. Yeah, this was about two years ago.

(Joel Beasley at 00:02:11) How did that transition happen? Was Sri like, hey, Rajeev, I've got something for you?

(Rajeev at 00:02:16) No, she's great. He was the CTO for about five years, I think, before me. And he took Atlassian from, you know, on-premise to the cloud and AWS and a whole bunch of transformations. And then, you know, as I came into this role, the founders, you know, Mike and Scott and the leadership, they were looking for someone who could help take Atlassian to world-class engineering. And what we mean by that is, Atlassian is a good solid company, and we have a great aspiration to be like a top tech company. Like, we think we could be in the top 10, and we could be a huge provider for knowledge workers. Our mission is to unleash the productivity of every team. And if you want to get to that, we believe that, hey, we need world-class engineering. We need our engineers to be the best in the world, like some of the other, you know, top tech companies. And so that was the goal coming in as I took this role on was like, okay, now what does that mean? What does world-class engineering mean? And as you break it down, there are three different buckets to it.

(Rajeev at 00:03:03) There is customer trust. How do you build products? How do you build software that earns the trust of your customers? And that is through security, through reliability, through customer support. Like, your products have to work well for your customers. The second pillar of world-class engineering is people, honestly. Like, you need the best engineers in the world. Okay? You need tech leads, you need amazing senior ICs and engineering leaders and architects, you know, and tech leads who can build great software. That's the second pillar, and I spend a lot of time on that. And then the third pillar is you want world-class engineering through developer productivity, great architecture, and things like that. And so that's the third component of it. And so we have done work in all three in the couple of years I've been here to kind of build up towards world-class engineering.

(Joel Beasley at 00:04:05) So world-class engineering is customer trust, people, and what's the third thing?

(Rajeev at 00:04:09) The third thing is basically developer productivity and great architecture, you know, which feeds into developer productivity. Right? If you don't have the right architecture, if you have modern code bases, then developers are more productive.

(Joel Beasley at 00:04:25) Do you guys do any of the vetting of people when they're looking to onboard engineers? I know you build a bunch of different tools. Do you do anything in the recruiting or validation of skills space?

(Rajeev at 00:04:37) Yeah, yeah. We actually revamped our whole interview process to look for the kinds of skills we would want in engineers, all the way from coding interviews to design interviews. For some senior leaders, we have a really interesting interview where we ask you, what's the hardest technical problem you worked on and, you know, explain to us what, how that worked and why it was interesting for you and what challenges you faced. I find it very instructive to hear from people about what they have built and how did they tackle the different challenges that came through. So we have done a bunch of things on the interview process. And then we've also done a bunch of things to onboard people when they join Atlassian. So we actually introduced—I was in Sydney about a month back, and we had our very first batch of Team Camp. We call it Team Camp. So you join Atlassian, whether it's out of school or an experienced hire, you don't just get into your job. You go to Team Camp for four to six weeks. You fix a few bugs. You learn the code base. You learn how to do things in Atlassian, and then you go to your team.

(Joel Beasley at 00:05:38) Oh, that's cool. Is that off-site or is that digital?

(Rajeev at 00:05:42) Partly, you know, it's a four to six week program of which we try to get you in on-site for a couple of weeks at least because, you know, when you're joining the company new, it's good to meet other people in person and build that connection. And so there's two weeks that's on-site and then the rest is virtual.

(Joel Beasley at 00:06:00) So your job is to take Atlassian to the next level with world-class engineering. And how long have you been at it, and how's it going?

(Rajeev at 00:06:09) I've been at it for two years. I've been at Atlassian for two years. And it'll be two years in a couple of months, so not quite. But—

(Joel Beasley at 00:06:16) Oh, you weren't a promotion from inside Atlassian?

(Rajeev at 00:06:19) No, no, no. I came from outside. Yeah, I was at Meta Facebook, you know, before that for five years. And then before Facebook, I was at Microsoft for many years. And I'm based in Seattle, and, you know, I was an intern on Windows 95 back in the day.

(Joel Beasley at 00:06:36) The most stable operating system. Yeah, that thing's still running. You'll walk by a screen in the middle of, you know, anywhere, and it'll be Windows 95 install running that display.

(Rajeev at 00:06:46) Right. Yeah, yeah.

(Joel Beasley at 00:06:48) So did you have hard metrics? I mean, this is a big job. Atlassian is a huge company, you know, to say, all right, we're going to hit this customer trust, people, and developer productivity. What type of KPIs were there or milestones? How do you refer to your goals internally? And then what were they to do this?

(Rajeev at 00:07:06) Yeah, and let's just take security and trust, you know, for example. When I came in, I looked at the scorecards and things like that, and it seemed like, you know, when you look at scorecards, it looks pretty green. I was like, okay, how do we benchmark this? I walked through a few things in terms of things I would expect, you know, in terms of protection at the perimeter or which engineers have access to production machines. You know, like, you don't want—you want to have a system where when you access a production machine, you go through some hoops and things, and you have good controls in place. And we didn't have a good set of things. And I was like, okay, we need to change who we benchmark against. And previously, we were benchmarking against other peer companies, and we changed it to benchmark against the top tech companies. Having come from some places like that, I knew some of the practices that, you know, companies strived towards in terms of scorecards. And so then our scorecard became more red, and then we came up with what we call the Top 10 Trust Commitments where, you know, we go to the whole SDLC, and we have scorecards around static code analysis and prevention in the SDLC cycle.

(Rajeev at 00:08:15) But then in production, how do we find vulnerabilities? How do we, in real time, detect problems? How do we make sure that access to production machines is secured? And so we have 10 trust commitments and scorecards against some of those things that I said, and we track that, you know, very, very closely month over month. That's in security and trust. On reliability, we look at the usual things. We look at SLAs for our enterprise customers. We look at SLOs. You know, if you build a microservice or a platform service, that's an internal service, it almost needs to have like five nines availability if you want to deliver like three nines or three nines five for an external facing, you know, interaction. So we take our top 10 interactions in our products, and we're like, okay, what is the SLA for each one of these things against our different customers? What are the SLOs of the internal services?

(Joel Beasley at 00:09:02) What's an SLO? I'm sorry to interrupt. I've never heard SLO before. What's SLO?

(Rajeev at 00:09:06) Yeah, it's more like a service level objective in the sense of, like, you have a microservice that's an internal thing that people use, and it has an internal objective in terms of we need this identity service, for example, to be five nines. Right? And it's not directly used by customers, but customer-facing products use that internal service. And so that internal service needs to be five nines if you want the customer-facing product to be three nine five.

(Joel Beasley at 00:09:38) Okay. All right. Yeah. And where are you getting—sorry, I'm a curious guy.

(Rajeev at 00:09:41) Sometimes it's called an SLI, like a service level indicator. But there is this concept of internal services that have a certain level of availability, and then there's external features that customers use that have what we call a published level of SLA. Right? So when we talk to customers and we say, hey, we give you three nines or we give you four nines, that's our published commitment to customers. That usually tends to be three nines or three nines five or four nines. But in order to deliver that, your internal stuff needs to be like five nines or six nines sometimes.

(Joel Beasley at 00:10:11) Correct, correct. And where do you get your scorecards, and how do you manage that?

(Rajeev at 00:10:16) So we build the scorecards through metrics. You know, we have logging. We use different tools and technologies, you know, for—

(Joel Beasley at 00:10:23) Like, if you were looking at the top 10 companies and are like, where are you—is there a company that looks at the top 10 and builds scorecards and says, hey, this company is doing it this way, this company is doing it that way? Or is that just all internal knowledge and you knowing people? Like, how does that get built?

(Rajeev at 00:10:41) So part of it is knowing what is great. You know, I have been at Facebook. I've been at Microsoft. Especially at Microsoft, we built Office 365. That was a team that I built. And—

(Joel Beasley at 00:10:52) That's pretty cool.

(Rajeev at 00:10:53) The infrastructure to take it to the cloud. So as we did cloud, if you go back 20 years and you look at all the companies that did cloud in the beginning, you know, this is all new stuff, so we kind of made up stuff along the way. But over time, we learned, you know, what works, what doesn't work, and what are the kinds of scorecards you need to measure, what are the things you need to measure, how do you have observability, how do you do logging, how do you take all that data and then build the scorecards from it. So there's a lot of best practice available out there. Some of it I have from my own experience. Some of it was already available internally in Atlassian, and we were able to build on that to build the right scorecards.

(Joel Beasley at 00:11:28) Mm-hmm. Yeah, I was just curious if there were people writing about—like, for example, Marty Cagan. I follow him for product. He's my spirit animal for product. I love his book. And he lists out like frameworks and ways of doing things, and I was just curious if there were people out there that were writing about these scorecards. Because I've heard it a couple times. I know there's like Security Scorecard, which is a software that does it for security specifically. But I didn't know if there was—so it's—but you answered the question, so thank you for that.

(Rajeev at 00:12:01) Yeah, yeah. I mean, some of this you can get from books and frameworks and thinkers, but I would say I myself, I'm more, I would say, partial to real-world experience, both mine as well as, you know, folks who have done it before for a decade or two. And there's a lot that's available just within the industry in terms of, okay, if this is the thing, what's the best way to go do that? Right?

(Joel Beasley at 00:12:25) So why Atlassian? Were you connected through friends? Like, how did you choose? You could go anywhere. You got Meta and Office 365, but you've done—you're a good name in the space with a lot of experience. Why Atlassian?

(Rajeev at 00:12:39) Yeah, look, I knew someone who happened to know the CEOs at Atlassian, and they were looking for a CTO. And, you know, sometimes you know someone you trust and they're like, hey, I think this is a good opportunity. Do you want to have a conversation? I was like, okay, you know, I'm happy to have a conversation. I talked to Mike Cannon-Brookes who is one of the CEOs, and I really enjoyed the conversation because he struck me, and he is an engineer. And like, the majority of the conversation, we were deep into technical code kind of thing. They were like, this is cool. This is just a cool conversation. You know? And then I was like, okay, I see the aspiration of where you want to take this company, and it's an iconic company because it builds Jira, which is something that's used by every technical team out there. And so as an engineer, for me, it was attractive to be in a company that builds software that's literally used by all engineers, all engineering teams in the world. So the leverage as an engineer—I've always been an engineer—was attractive.

(Rajeev at 00:13:32) And then the aspiration to then make knowledge workers more productive, all kinds of knowledge workers, not just engineers, but also marketing teams, HR teams, you know, non-tech teams, that's our aspiration is all teams should be able to use Jira to be more productive. If you're building a rocket ship to go to Mars, you should be able to use Jira to track your bug, you know, your sort of project burn down if you want. Right? And so that was attractive. And having been at big companies, it was interesting for me to be at a smaller company that's trying to become bigger for two reasons. One is I'm coming in at a leadership position, and I can help guide the whole company towards that, having been in other places that were bigger, so I know what that scale looks like. But I also want the journey to be—you know, when a company becomes big, I've seen bad things happen too. I've been through the down years at Microsoft. I've been through some down times at Facebook. So I've seen some things that go wrong when a company becomes bigger. And for me, it was attractive to be able to at least help, hey, avoid these bad movies. Okay, don't do these things. Right?

(Rajeev at 00:14:36) Lean into these playbooks. They will help you do better things. And so to bring my experience to a company that's trying to grow bigger was interesting.

(Joel Beasley at 00:14:45) When I started seeing those videos pop up and go viral of those project manager people who were, oh, I show up at ten. I drink coffee. I lounge outside over at Meta.

(Joel Beasley at 00:14:58) When I saw that, I was like, they're gonna have some problems. Luckily, Zuckerberg responded pretty quickly. I think he did the best — there's people that completely miss PR opportunities, and I pay attention to it now that I'm in media. But I think he did a great job. He came out and was just like, hey, we're leaning up. We're firming up. We're not doing that anymore. And that's great. So I think he handled that.

(Rajeev at 00:15:18) No, I think at both places I've been, you know, Meta as well as Microsoft, I've seen the great times. I've seen things go down. I've seen things come back up again. And so those are good lessons, and I hope to be able to guide Atlassian through some of those kinds of times.

(Joel Beasley at 00:15:36) So a lot of the people that listen are trying to advance in their technology career. Maybe they're moving from IC to team lead or team lead to leader of leaders or to VP or CTO, CIO, CXO type deals. Transitionary periods. Right? How do you feel about individuals who are individual contributors on a team, and they think that they might want to be a leader of a team?

(Rajeev at 00:16:02) That's a topic that's very close to my heart because when I started out as an engineer at Microsoft, I remained an IC, senior IC for as long as I could. In fact, I explicitly avoided being a manager. They tried to make me a manager many times, and I was like, no, I don't wanna be a manager. Partly, I got some really good advice from senior leaders who said, early in your career, just write a lot of code, go deep, get technical, understand things really well. And at some point you can switch to being a manager if you want to, or you can continue on the IC path. In fact, Microsoft and many other companies have a path where you could become a distinguished engineer or technical fellow or something like that. And you don't ever have to be a manager if you don't want to. In fact, going back about thirty years in tech, we made the mistake of making our best coders managers. And we quickly found out that that doesn't work. You could have people who are awesome coders, but really bad managers. And over time in tech, I think we have figured out a better way to go do this. But in a good company, you would have a great path both as an IC and a manager. You won't have to decide that in order to grow in my career, I have to manage people, I have to become a manager. That's not a good thing for either for the team being managed or for the person who's trying to advance their career. So that's what we're doing at Atlassian as well. In the last couple of years, we actually — I've hired two distinguished engineers to Atlassian who are really setting a role model for this is what a senior IC looks like in engineering. And we have really made it very clear that you could grow as an IC or a manager and both paths can take you as high up in the company as you want. And so to me, that's very important that we have those two paths available so that people can choose. And if you really want to be a manager, we actually make you — the first time you become a manager, we put you through an interview process to really understand what is your true motivation. And one of the things I do when I speak to candidates is I try to convince them to not become a manager. Like, why do you wanna be a manager? You're writing code. You're having fun. Don't go down this path. And but then you get into the true motivation of why they wanna be a manager. And usually, it comes down to, I really enjoy coaching people. I really enjoy seeing people advance and helping them out. You know, you get the right motivations, and then you're like, okay. This person can actually be an amazing manager. Because we don't want average managers. We want really amazing managers and leaders or really amazing ICs. We don't want people who are not doing a good job either.

(Joel Beasley at 00:18:29) Yeah. So that would be a positive thing that you're looking for is someone who wants to coach people and grow them, and that's what you're looking for in your managers?

(Rajeev at 00:18:40) Yeah. And you gotta be willing to do whatever it takes to help the team succeed. And at the same time, you also want people who are like — I still wanna be technical. I'll pick up the bug that no one wants to fix or that nobody is able to fix. So I continue to stay as hands on as I can, even while helping guide my team and so on. Because we don't want the other thing. We don't want people to become pure managers where they lose touch. We all joined to write code and to build things, and so you don't wanna lose touch with that.

(Joel Beasley at 00:19:10) What's a — if I'm an individual contributor and I tell you that I wanna be a manager, what's something that I would say that would show you that I'm not right to be a manager?

(Rajeev at 00:19:20) Oh, I mean, you know, usually it's not may not be said directly, but it's sort of like, I feel like I can advance in my career faster by being a manager, or I will have more control if I'm a manager. Those are words that are flags for me because we expect you to be able to influence things whether you're a manager or not. And if you think that just because you're a manager, you can control what people do, then that tends to be a flag for me because engineering is a very tenuous thing where you want to have bottoms up energy, and you don't want managers who are trying to drive things too top down. In fact, I believe in what's called selfless leadership or servant leadership where it's not that they work for you. You work for them. If somebody reports to you, you are working for them, and you want managers to have that kind of feeling. But if they're wanting to do it because of control, then that's not the right flag.

(Joel Beasley at 00:20:18) Are you a parent?

(Rajeev at 00:20:28) Yes. I have three kids.

(Joel Beasley at 00:20:29) How has being a parent helped you be a manager and a leader of others?

(Rajeev at 00:20:31) More than anything else. You know, you want the best for your kids. At the same time, you don't wanna micromanage them. So you find that balance between telling them what is right and wrong, but also helping them find their own way and not being too prescriptive about things. And because they are your kids and you want the best for them, as parents, we always sometimes are doing too much parenting. And then, of course, at different ages, they'll tell you to back off and you'll learn so much at each stage of their life as well. So, yeah, I can say that I'm definitely a better manager leader because of being there, for sure.

(Joel Beasley at 00:21:08) What's the age ranges of your kids?

(Rajeev at 00:21:10) My older one is — he's graduated from college a year back, so he's just in a new job. And my younger one, she's in college.

(Joel Beasley at 00:21:19) Oh, wow. Dude, you don't look that old. How old are you? This is just you and me hanging out.

(Rajeev at 00:21:25) Let's put it this way. I told you. I gave you a hint. I was an intern on Windows '95.

(Joel Beasley at 00:21:29) Yeah. 1995. Yeah. That's right. Alright. Alright. Yeah. So I'm 36, and mine are ages two to seven. So I'm — yeah, I'm in the — you know where I'm at.

(Rajeev at 00:21:43) They say the terrible twos. Two can be a tough age, but, you know, it depends on each kid.

(Joel Beasley at 00:21:50) It does.

(Rajeev at 00:21:51) But yeah. I say, using computer science analogies, I'm like, in the beginning, it's more like IO problems and then the shape of the problems change. But they are always your kids and they get more degrees of freedom. So the shape of the problem changes. Like, in the beginning, it's more basic stuff. And then over time, eventually, one of your kids is gonna be driving a car, and that's gonna be a big day.

(Joel Beasley at 00:22:21) Yeah. I've got a lot of friends who have a lot of kids, and I've got one right now. One of my good friends has seven kids between the ages of five and 16. And so he's got one at every different developmental level. I was like, you're a champion, dude. That's a lot of kids.

(Rajeev at 00:22:38) Yeah.

(Joel Beasley at 00:22:40) So alright. Yeah. So management for me was an interesting thing because if you were to talk to me back before I was managing people and I was just an individual contributor, I'd be like, I'm gonna be doing this forever. This thing brings me maximum joy. I love solving a problem. I love the explicitness of dealing with these problems over dealing with human problems. You know, humans don't have output logs telling you everything what's going on. And so that's what I would have said. Until I was put into the position where I was running a — I built a piece of software that was doing well, and I needed to hire people to help me build that software because I was making money and I needed it just had to grow from single developer project to multi developer project. And then I started working on people and then I found that people is a fascinating problem. It is really complicated. But it's also, you know, when I see this move and hear people like you talk about moving from IC to first time leader, it feels like it's a progression, but it's not. It's an entirely different job. It's like saying you're going from the ice cream factory to running marathons. It's a completely different set of skills, completely different set of principles, books you need to read, things you need to understand. It's just so totally different. And I found it very challenging. And so, to me, it was exciting. And so I figured out how to get good at doing that. Then I realized it wasn't for control as much as it was for impact. I thought it's cool that I can now direct swarms of engineers towards an outcome that no individual can achieve on their own. And to me, that was cool. And so that mixed with working with people to achieve some sort of higher level outcome became the new thing I was addicted to. I was like, that's the thing.

(Rajeev at 00:24:30) If I were interviewing you right now for a manager position, I would hire you because that's exactly the answer I wanna hear.

(Joel Beasley at 00:24:37) Well, thank you. I'll accept the job. Connect me with HR. No. But you know.

(Rajeev at 00:24:43) That was my journey too. I was an accidental manager. In fact, what happened is, like I told you, I tried to avoid being a manager. And then at one performance cycle, my manager came to me and said, you are doing the performance review for these three engineers. I'm like, wait. Hold on. I never said I'm a manager. You never told me this. I'm not doing this. And they're like, no, Rajeev. We really need you to do it. We basically think of you as a manager for these three people and do it. And so I spent the whole night, like, big part of the night at least, going through the code of every one of them with a different lens. So far, they were colleagues writing code. We're writing code together. And now I'm like, okay. I need to go and assess the code. So let me understand what's the design? How is the code laid out? And I came up with some ratings, and I forgot about it. But over time, as I did take on manager responsibilities, I found that there are these different scaling points. You know, when you go from managing three engineers to four engineers, I found my core production came down. When I went from four to seven, I was doing more code reviews. When I went from seven to 22, I was not even familiar with every line of code. And so at every sort of number of engineers, I had to change the playbook of what I did as a manager, you know, going all the way from line level to skip level to director level. It just changes. So to your example of going from an ice cream factory to a marathon, even as you go up in leadership levels, it changes completely. There are these breaking points. And if you don't realize that you've gone past a particular scale point and you're trying to run the old playbook, you'll fail. It literally won't work.

(Joel Beasley at 00:26:25) Yeah. It — to your point, when I see people ask — people write in from the show or they ask questions online that I see. And whenever I see somebody asking the question about, this is the job title, what are the responsibilities? I'm trying to determine what I need to be doing based off of the job title. And I'm like, well, it's gonna be different inside of every different business because every different business. So you have to understand — the skill is to understand the needs of the business and what you're able — how you're able to help with the business achieve those outcomes that it needs. And that's gonna change constantly.

(Rajeev at 00:27:00) That's right. That's right. Yeah. And you have to be thoughtful about it. You have to think about, what is the job I'm doing for the company? What are the outcomes you wanna get? And, therefore, what is the playbook I'm gonna use to go do it? One of the principles I have is, when I look at things, I think about problems. I'm like, should I manage this through people or should I manage this through process? And by that, what I mean is, in some cases, I have a problem, let's say security. And I might go and hire or I might have a person who is really amazing at security, and I know this person well. I trust the person. So I manage through that person. I keep in touch by talking to that person. They are going off managing different things, but I keep in touch with that person. That's how I manage it. In some other areas, I may not have a person. It might be two, three different leaders. So then I create a process. I'm like, okay, let's meet once a week or twice a week, and let's go through these different things so that I understand it myself. I don't have a single person I can rely on, but I have a process or a framework through which I can manage that problem.

(Joel Beasley at 00:28:03) Oh, that's good. And these principles, these are the things that help you work at these higher levels?

(Rajeev at 00:28:11) Yeah. Yeah. You sort of have to evolve a rubric of things. So manage through people or manage through process is one such thing. The other one is a physics example I learned from a leader at Microsoft. And I call it conduction, convection, radiation, which is, you know, conduction is when you touch a person directly or you interact with a person directly, that's conduction. When you do it through another person, it's convection. When you do it by just trying to reach a bunch of people through a blog post that's radiation. So if you're managing 10 people or a 100 people or a thousand people, you can't just say, I have a thousand person org. I'm just gonna manage it through my direct reports. It's not going to work. You have to find a way to get to know all thousand people. You can't literally get to know all thousand people because there's not enough time in the day, but you gotta find ways to open yourself up so that you meet as many people as you can. You interact with people through blog posts. You do skip levels. You do group meetings. You repeat your message so that everyone understands it. But you have to take the effort as a leader to really actually impact. And through conduction, convection, or radiation, you touch everybody in the organization.

(Joel Beasley at 00:29:21) By the way, I love that. That's brilliant. How do you find a way to get to know a thousand people?

(Rajeev at 00:29:27) In my case, in Atlassian, I have about five, six thousand people overall in my organization in terms of number of engineers. I traveled a lot. I went to Sydney a lot. I have a lot of people in Sydney. I went to India.

(Rajeev at 00:29:39) I went to, you know, I'm in Seattle. I have come to Bay Area. When I am in these locations, I set up literally a lot of meetings with small groups of people. You know, you can do a town hall or you can do a big event, and you can do Q&A and things like that. And that's a good way.

(Rajeev at 00:29:58) But I find more energy from meeting groups of ten, fifteen engineers in a room and we take a component and, okay, deep dive. Let's deep dive with the component. Literally, we pull up some code. We go to the whiteboard. We talk through things, and I get a really good sense of I learn things, and then I give some time for Q&A so that they can ask me questions.

(Rajeev at 00:30:16) And I come away from those meetings getting to know some people, but also getting to know about the area. And I do a lot of those, you know, through the year or as I travel and I visit different locations. That tells me a lot about, you know, different folks. And just as a leader, you have to be accessible. Like, if somebody Slacks you or messages you, being open to that and being able to connect with different people in the organization and being accessible, I think, is an important thing.

(Rajeev at 00:30:44) You know, as an engineer growing up, I remember how my leaders used to be and who was accessible and who was not. And I always felt like if I ever got into a leadership position, I wanted to be connected with engineers and really connected with the reality on the ground, as opposed to being high up there in abstract land.

(Joel Beasley at 00:31:03) And you also realize that you have to make an effort because they're less mature in their career often, you know, where there are people at all various sizes of the shapes and steps of their career. But with the newer people, you have to make more of an effort to just constantly remind them because they by default, when you're newer in your career, you look up at the person at the top and it's scary almost. And you're like,

(Rajeev at 00:31:26) Oh, I

(Joel Beasley at 00:31:27) don't know. They're busy. They're busy. They won't, you know, talk with me. And so you have to be like, hey. The door's open. We're here.

(Rajeev at 00:31:32) That's right. Yeah. Yeah. Yeah. Yeah.

(Joel Beasley at 00:31:34) Now when you're doing the Q&As, so when you're going around and doing the Q&As, obviously, you're getting questions about the specific components you're talking about. But what type of career growth or leadership questions do you hear a lot?

(Rajeev at 00:31:49) Yeah. I think some of it is, what are your leadership principles or, you know, what are your values? I hear the values question a lot nowadays. You know? Honestly, when I started as an engineer at Microsoft, there was not so much about values and mission and vision statements and so on. My values are very simple. Write code, make sure it has no bugs, check it in, and then don't break the build. You know? But we didn't have as many value posters on the board, on the walls and things like that. That seems like a more recent thing.

(Rajeev at 00:32:22) Like, in the last, you know, ten, fifteen years, I see that a lot, and I have people ask me about it. And so I talk a little bit about my values as I have learned them over time. And then people ask about, you know, projects that failed or how you learn from those kinds of things, you know, things like that. And they're also interested in sort of, how did you scale your leadership over time? You know, things like some of the things I talked about just now.

(Joel Beasley at 00:32:49) What have you noticed about this next generation coming into the workforce? Have you gotten to spend any time with them?

(Rajeev at 00:32:57) Yeah. Yeah. Two ways. One is my own kids are, you know, in that age and stage, and I learn a lot from them. And then I try to spend time. I always remember my time as an intern, as a fresh grad. And so I try to meet up with folks, you know, who are in that stage of their career as well. And I find first of all, I'm super impressed. Like, I feel like when I was their age, I didn't, you know, have that level of curiosity. And the thing that strikes me about them is they're very mission oriented. Like, they actually care a lot more about what am I working on, what is the mission for it, how is it gonna change the world, how can I make the world a better place?

(Rajeev at 00:33:40) I don't know. Maybe it was me or I don't know what, but when I started out thirty years back, it was more about, okay. Can I work for a good company? What will I learn? And, you know, I didn't think as much about mission as much, I guess. You know? And maybe it's me, but, you know, maybe it's my generation, but I just feel like the folks coming in now are much sharper about what they work on.

(Joel Beasley at 00:34:03) Why do you think that is?

(Rajeev at 00:34:06) Maybe it's to do, you know, with how the world has evolved and, you know, that sort of thing. I don't know.

(Joel Beasley at 00:34:14) Well, we become more connected than ever.

(Rajeev at 00:34:15) Yeah.

(Joel Beasley at 00:34:16) I mean, three years

(Rajeev at 00:34:17) That's fine.

(Joel Beasley at 00:34:17) I'm 36. I mean

(Rajeev at 00:34:18) Yeah. Yeah.

(Joel Beasley at 00:34:19) I was the last, so I have a brother who's three years older. He's a doctor. And I have a sister who's three years younger. She's a physics teacher. And I was the last generation that was, you get locked out of the house when you come home from school, you drink from the water hose if you're thirsty, and you ride your bikes in the woods until it's dark, and then you get home.

(Rajeev at 00:34:40) Right.

(Joel Beasley at 00:34:40) And then somehow within a two year period, you were a criminal if you let your kids walk to the park by themselves. Yeah. Yeah.

(Rajeev at 00:34:49) I'm like,

(Joel Beasley at 00:34:49) I don't know what happened.

(Rajeev at 00:34:50) Yeah. Yeah.

(Joel Beasley at 00:34:50) Yeah. But at least I got in there, and then I got the, you know, cell phones in high school. We got cell phones and, you know, all that good stuff. So I kinda went from no technology to early adopter to through on that last generation. It's been really interesting to see this next generation come up. But, yeah, we get more connected. Yeah. And then life got better. Like, we are wealthier and healthier.

(Joel Beasley at 00:35:18) Well, I don't know about healthier in the states, but I am. Yeah. Yeah. Well, we're healthier priorities change. Right? It's like Maslow's hierarchy of needs. And so

(Rajeev at 00:35:29) That's right.

(Joel Beasley at 00:35:30) The economy got so good and there was so much wealth built that people are like, well, now I'm just gonna care about the mission. Maybe that's the issue.

(Rajeev at 00:35:37) You're right. You're right. That's a good that is exactly right. In my days, I think we cared more about, okay, you got a good job. Hey, you got a job at Microsoft. Great. And even the sort of culture was one of tough love. Like, nobody really gave you a lot of praise for doing good stuff. And if you did bad stuff, like, you broke the build, you got, you know, on the wall of shame, so to speak. Right?

(Rajeev at 00:36:00) So, you know, whereas now, I think folks expect more positive reinforcement and

(Joel Beasley at 00:36:10) I could imagine a wall. You leave this conversation. You're like, alright. I just talked to my manager and CTO guy. We're gonna create a wall of shame.

(Joel Beasley at 00:36:20) So, hey, I wanna be respectful of your time. I also wanted to touch base really quickly on how, and there's gonna be things you could talk about, things you can't. So if it's a can't, we can just move on. But I was curious with this rise of the large language models in the past year

(Rajeev at 00:36:38) Mhmm.

(Joel Beasley at 00:36:39) Yeah. Has that made its way into any Atlassian products?

(Rajeev at 00:36:41) Yeah. Yeah. Yeah. Absolutely. In fact, last year, we announced our first wave of AI products. We call it Atlassian Intelligence. We have an annual event in Vegas. We call it Team. You know, last year was Team '23. This year was Team '24. But at Team '23, we announced Atlassian Intelligence. We announced a set of features, and eventually, we shipped it in later part of 2023 to all our customers where you can go into a Confluence document and it'll auto summarize for you, or you basically get an AI mate. It's like a teammate who's a virtual member of your team who can answer questions for you in context on projects that you're working on and so on.

(Joel Beasley at 00:37:20) Can I hook into Compass?

(Rajeev at 00:37:22) Yeah. Yeah. Absolutely. We can. We're working on it, but we have it with our main products like Confluence and Jira. And then we are also working on things where you can do natural, we have a query language called JQL for Jira Query Language, and you can express natural language statements, and it'll do the translation for you to JQL. So we built some of those features, and we actually built it on OpenAI. So we have a partnership with OpenAI where we use their LLM to build some of these things. But at the same time, we are experimenting with all the other LLMs out there because it's fascinating. I mean, it's like the LLM wars.

(Rajeev at 00:38:00) And from our perspective in Atlassian, we are in the business of building features for our customers, unleashing the data for them, and we stand to benefit from whichever LLM succeeds, you know, in terms of cost or quality and those kinds of things.

(Joel Beasley at 00:38:15) It's like the browser wars all over again.

(Rajeev at 00:38:18) Exactly. Or the cloud, you know, the cloud wars, right, with AWS, Azure, GCP, and so on. I think the big providers will commoditize LLMs to the point where it'll get faster, cheaper, better quality. And companies like Atlassian stand to benefit because we are gonna use the best of breed to go and build the best features for our customers.

(Joel Beasley at 00:38:41) I just saw the other day that they're pushing to get models greater than 20 gigs being registered with the US government.

(Rajeev at 00:38:48) Oh, wow. Yeah. Yeah.

(Joel Beasley at 00:38:49) So I mean, it was an executive order, but I don't know how that'll work. I don't, I'm not a constitutional attorney. Yeah. Yeah. I'm not a political person either.

(Joel Beasley at 00:39:02) Yeah. But I also am excited for the decentralized. There are some blockchain projects that are attempting to do decentralized AI models so that they're always, because the cat's kinda out of the bag. Right? The model's already out there. So

(Rajeev at 00:39:17) Yeah. But the models are changing too. Right? So people are doing custom models. People are doing, you know, all kinds of new models. Every day, there's an announcement by some company about some new models. So I think it's interesting. As an engineer, as a technologist, to me, it's super fun and fascinating that we have all this innovation going on.

(Joel Beasley at 00:39:36) Oh, me too. I am very excited about it. And did Elon open source his Grok yet or no? I don't know.

(Rajeev at 00:39:45) I haven't kept up with that.

(Joel Beasley at 00:39:46) I haven't played with that at all. Have you played with it at all?

(Rajeev at 00:39:48) No. I haven't. No.

(Joel Beasley at 00:39:50) I think that's what it was called. I just saw the headlines. So, you know, there's only so much headlines you can read. You gotta do your work.

(Joel Beasley at 00:40:00) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn or send me an email [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.