In this edition of the Datadog Engineering Spotlight, Rosa from the Community team sat down (virtually) with Jeromy Carriere. He’s SVP, Product Engineering, leading engineering for all of Datadog’s product areas including Infrastructure Monitoring, Log Management, Application Performance Monitoring, Security, Service Management, and User Experience Monitoring. We talk about his relationship with academia, learning from his mistakes, driving the first cloud monitoring offerings at Google, and the future of observability.
This interview has been edited for clarity and length.
What’s your day-to-day like?
It varies to a degree based on a few different cycles: The planning cycle at the beginning of a quarter, where I try to look thematically for opportunities to connect things and get more leverage out of collaboration. And the execution cycle, when I look in a cross-cutting way for opportunities to support teams by unblocking them where they need more resources, or unsticking them on some decision-making angle.
Another cycle is the people side of the organization—not just the performance of my direct reports, but also supporting the broader organization in the performance management process, which any mature engineering organization needs to be rigorous and consistent about. One critical thing in my day-to-day is around the function of Engineering—how does Engineering work? This is sometimes dealing with process definition, refinement, or improvement. Other times, it’s about how we hire and manage performance, so it’s more at the meta level.
One critical thing in my day-to-day is around the function of Engineering—how does Engineering work?
I also spend time reading design docs and code, observing incidents and post mortems, and keeping an awareness of how the organization is executing across the entire stack. I recognize this is a long answer, but these things all demand different amounts of attention at different times. It’s a matter of keeping them in balance.
How would you describe your journey to this position?
In 2014 when I worked at Google, we were early in the journey of Google Cloud Platform, and didn’t have any Datadog-like capabilities for our public cloud customers. It substantially compromised our ability to build a viable business. So, my primary job was to produce a cloud monitoring offering. It was during this time that I met Alexis and Amit. We had a great engagement, and it’s where I first had an early awareness of Datadog. At Google, we eventually acquired a company and built our own cloud monitoring offering. After some time, I moved on to Facebook, where I did some observability work, along with a variety of other things.
When I decided that I was ready to move on from Facebook, I reconnected with Datadog and made two realizations: one was that while at Google, I had developed a lot of passion and excitement for the observability space, which comes from the opportunity to impact the productivity of developers and engineers and all the associated roles and responsibilities by doing what Datadog does really well.
I was watching how the company kept innovating and delivering products with sustained velocity—and I use that word intentionally. I don’t just mean moving fast, but I mean moving fast with direction and an intentional strategy.
The second realization that led me to Datadog was that I was watching how the company kept innovating and delivering products with sustained velocity—and I use that word intentionally. I don’t just mean moving fast, but I mean moving fast with direction and an intentional strategy. That got me really excited. Like, how does this company that has been around for 10 years maintain this product delivery velocity for this length of time? That was a strong attractor for me: to learn how I could enable the organization to maintain this velocity as it proceeds to scale and take further steps in its growth.
What are the lessons you’ve learned along the way?
I’ve come to the realization that I only learn well when I make mistakes. I’m not saying I always only make a mistake once, but you don’t want to repeat the ones that you’ve already made. I joke that everyone should always be making new mistakes. You want to learn by making and owning your mistakes. Learn from them and grow.
I think that mistakes I’ve made in the past are the strongest lessons that I bring to Datadog. Early in my career, I thought that I had to be very directive in order to get outcomes, but this approach compromised my team’s ability to be creative and feel in control of their own direction. At other points, I was too far removed and didn’t give teams enough support to fulfill their commitments. I now try to strike a balance between enabling teams to execute with as much autonomy as possible, and holding them accountable for the outcomes that they commit to. That’s a lesson that I’m not saying I’ve gotten exactly right, but it is probably the single largest lesson I think is relevant.
I joke that everyone should always be making new mistakes.
The other lesson I’ve learned is about helping people to grow and develop in their roles. There’s an implication that people know what it is that they want and where it is that they’re going with their careers. But I know from my own experience that it took me a long time to come to a place where I could conclude that I would be most impactful as a manager. After flipping back and forth between being an individual contributor and a manager, I decided that being a manager was the path, but at the same time, I knew I was not good at it. I wanted to get better at it in order to have a positive impact.
I don’t expect anyone to know their five-year plan or the steps they want to take in their career. I know that’s unrealistic from my own experience. But you should know what gives you satisfaction and what you feel good about doing so you can find opportunities to feel that way and therefore bring your best. In my opinion, people don’t bring their best to something that they don’t feel satisfied by. That’s one that took me a long time to learn.
What are tactics you use to help people figure out the next career step?
This is easy to say and hard to do, which is that you’ve got to take a step back. We all get very heads-down and we know how to pursue the thing that is in front of us. But we’ve got to sometimes look to the sides and say, well, what am I missing? What are other things that are around that I should be considering? So you’ve got to give yourself space and freedom and license to do that. And that is an explicit, intentional step everyone needs to take—to give yourself some time to make some observations of what makes you feel satisfied.
Tell me about your experience with the co-op program at the University of Waterloo.
I give the Waterloo co-op program a lot of credit for setting me on a successful path. And that comes from the opportunity to get an early appreciation of what it means to be a professional software developer. As strong as the academic program is at Waterloo—and I believe it’s truly strong—the real value is the combination of that program with the regular cadence of high quality placements with employers, where I was given real work to do.
And we do this at Datadog as well—our interns do some of the best work. That opportunity is really key to success. Lots of schools do this—Waterloo does it well from a practice perspective, and there’s a strong alignment there.
Our interns do some of the best work.
I’ve recently become an Adjunct Assistant Professor at Waterloo and will co-teach a class in the software engineering program in a guest lecturer role. Part of my personal goal there is to scratch that itch to be connected to the academic world.
Besides software engineering, what things are you passionate about?
I’ve got a few hobbies. I like rebuilding, reconstructing, and repairing old computers—vintage, retro computers from my childhood—from 80s and 90s, but mostly 80s.
My other hobby is driving race cars. I came to this much later in life than a lot of people—only eight or nine years ago. When I have time, I like to drive on the track. Lots of us as kids dreamed of being a race car driver, but I know now for sure I would never have been any good at it at the professional level. But it is a passion.
When did your hobby of restoring old computers start?
During the pandemic, I started watching YouTube channels of guys doing this—it’s always guys, I’m not saying that unintentionally. Although, there are a couple of women that are well known in the community. Jeri Ellsworth is one.
It’s only been a few years, but I have within arms reach, [counts] six retro computers. And then I have this one, an Altair 8080 board [shows board], this a reproduction of one of the first “personal” computers, from the late 70s—I’m still building it—it’s just the bare circuit board with a few components on it. I like to get reproductions and rebuild them or find an old one and try to repair it. Sometimes that works, sometimes it doesn’t.
What other things do you enjoy doing outside of work?
I like to cook. And unlike my wife, who’s an intuitive cook, I’m a recipe follower. I like the predictability of it: you know when you start, you know when you’re done, and you’re following a plan. Most of my day-to-day work involves coming up with a plan, and you never really know when you’re done. So I like that aspect of it. And then, hopefully, you get good food at the end.
And I love animals. When I was at Facebook, I did a team all-hands with the usual presentation type stuff, but then I had a farmer from a rescue sanctuary farm in upstate New York join and she toured us around on video to see a bunch of animals on the farm—goats, chickens, turkeys—which was a lot of fun for everybody.
How did your upbringing influence how you work today?
I was well supported by my teachers and my parents as I developed a passion for computers. Something I’m doing with the University of Waterloo, separate from teaching, is supporting a program for childhood education as part of the Centre for Education in Mathematics and Computing. It fosters attention for this type of curriculum and support in early elementary school children.
I was so well supported that it set me on a path that I wouldn’t have been able to follow without that early support. So, my upbringing was just very enabling of the things that I loved, and maybe that spoiled me, but it gave me a lot of opportunities to pursue what I wanted to do.
What is your favorite part of your job?
That’s not an easy question because there’s lots of favorite things. I think, probably, the one that I would have to go to is the people. That sounds like a throwaway answer that everyone’s going to give, but it really is true. My team, my peers, my partners, the broader organization, and our executive leadership. I really like the people I work with.
That is a baseline for enjoying the rest of my work, because otherwise you can’t be satisfied. No one would say, I want to do this and I’ll work with a bunch of people I don’t like working with, and it’ll be okay.
What do you think of how the observability space has evolved and where it’s heading?
I have a lot of passion for the space because I really believe that if done well, products, tools, and services in the observability space can materially improve both the lives of engineers and the quality of the products that they build and serve their customers, so I think there’s a huge opportunity to have an impact.
The space has evolved enormously. Back in mid-2010s, I was at Google, and we were just getting started at Google Cloud; Amazon had been around for close to 10 years at scale at that point, and it was the early days of Datadog as well. It’s amazing to think back at how minimal the tools were and how far we’ve come both in the commercial realm and in open source.
You see pretty interesting things emerging, like OpenTelemetry, which is the open standard for observability data (metrics, logs, traces, and so forth). I think OTel is reflective of a maturation of the space in general. I personally look at OpenTelemetry as a huge opportunity for us. Certainly it’s got some complexity in terms of how we approach it and how we bring it into our narrative and our technology portfolio, but it’s also a chance for us to push forward the reach of observability tools in the industry. And our customers are asking for it, so it’s an example of an evolution that is really reflective of a maturity in the industry.
[OpenTelemetry] has some complexity in terms of how we approach it and how we bring it into our narrative and our technology portfolio, but it’s also a chance for us to push forward the reach of observability tools in the industry.
The other thing that is top of mind for everybody right now is AI and Large Language Models. We announced a ton of stuff at Dash around LLMs, both the application of LLMs and the observability of LLMs and LLM infrastructure. As is true for many domains, the application of LLM to observability is an interesting opportunity for us. We’ve launched Bits AI, which provides the powerful ability to have a natural language interaction with Datadog.
One of my favorite recent features in the product is automatic incident summarization. If you join a Slack channel for an ongoing or resolved incident, you automatically get a summary of the incident so far, generated by OpenAI. It’s so impactful! A small feature that hides an enormous amount of technical complexity and power. But it is such a material improvement. I’m used to joining an incident kind of late because it’s been escalated to me, and I need to jump in and understand the state. Before this feature, I had two choices: I could either bug the incident commander to give me a summary, which is a waste of their time because they are busy remediating the incident; or, more likely, I could scroll back through the whole Slack history and try to build my own synthesis of what’s happened. LLMs do a great job of that kind of summarization.
So, LLMs represent another evolutionary stage for the industry, and the world at large, and has notable opportunities for impact in the observability space.
What do you think is exemplary about Datadog?
The rate at which Datadog continues to innovate and continues to bring new capabilities into our product portfolio really sets us apart. It’s one of the reasons I came to Datadog: to participate in this and learn from it, and hopefully contribute to this product delivery machine that is not just doing more of the same—turning the crank—but continuously producing really, really useful and value-adding capabilities for our customers. That’s sort of a meta thing that I think is very special about Datadog.
The other thing, slightly more technical but also quite meta, is the platform nature of Datadog. It’s a thing we sell on and talk about as a special thing about Datadog, and it’s true. Our products weave together into a whole and mutually reinforce one another. Being able to, for example, navigate between different parts of a product seamlessly, maintaining contextual awareness—that’s a thing that you can struggle with. I struggled with this at Google and Facebook. At Google, we never solved this problem for the Google Cloud monitoring offering, and we had clear boundaries between metrics, logs, and traces. There were some light integrations between them but they weren’t smooth and that meant that the user friction was unnecessarily high. So that’s an example of a thing I think is special at Datadog.
Our products weave together into a whole and mutually reinforce one another.
Last thing I’ll mention is more of a transitional thing for Datadog, and that’s our move from being purely a passive observer, which is historically where we focused, to becoming more action-oriented on behalf of our customers. So from just observing and giving the customer data, to offering insights and recommendations, and then to offering the opportunity for Datadog to take action on their behalf, executing workflows that adjust security or infrastructure configurations. That’s really a sea change in the way Datadog is positioned with respect to our customers’ expectations, and that, I think, is very powerful.
How does it feel to have your own Wikipedia page?
It’s fun because my kids’ classmates discovered that I—their friend’s dad—had a Wikipedia page. But it’s also kind of embarrassing because I don’t have any real claim that I should have a Wikipedia page and at some point someone did try to delete it. But it exists because a good friend of mine and co-founder in my first startup company, Quack.com, created it. His name is Steve Woods and he’s maintained it from the beginning. I don’t know if anybody else has even contributed to it. I don’t ever really look at it. Occasionally I go to see if it still exists, but yeah, it’s kind of fun and weird.
Many thanks to Jeromy for sitting down to share his experience and insights with us! If working with people like Jeromy who are passionate about observability and who foster innovation and intentional career growth, check out our Careers page and join the pack!
