August 1, 2018

The Engineering Hangout with Vlad Cretu: Building products for engineers

The Engineering Hangout is a podcast where engineers hangout and talk shop. Hosted by Chris Reeves, the show is brought to you by Templarbit. You can subscribe on iTunes, Breaker or anywhere podcasts are found.

In this episode our CEO and Co-Founder, Bjoern Zinssmeister interviews Vlad Cretu, VP of Infrastructure at Sentry.

Sentry provides engineers with Open-source error tracking that helps monitor and fix crashes in real time.

Sentry launched in 2008 as django-db-log on Github and renamed to Sentry in 2010. Today the product is available on many platforms including .Net, Rails, Go, and iOS. Sentry is based in San Francisco and recently closed a $16M Series B round in funding.



Templarbit engineering hangout vlad cretu bjoern zinssmeister Vlad Cretu talking to Bjoern Zinssmeister at the Templarbit office in San Francisco.

“Overall there needs to be a very explicit way around what we expect out of the Engineering Management function. And it needs to be communicated and understood by everyone in the organization…”








Vlad’s career as an engineer spans over 15 years including time at Microsoft, McAfee and DocuSign. Before Sentry, Vlad was the Head of Engineering for the Hipchat Server at Atlassian.

During the episode you’ll hear:

Learn more about Vlad Cretu at Sentry.io.

Learn more about Templarbit with our free 14 day trial.


Full episode transcription below:

[electronic keyboard music fades in]

00:00:03:09 - 00:00:08:15 | Vlad

I would even go out to say that every college hire should work at Microsoft for six years.

00:00:08:15 - 00:00:08:20 | Bjoern

Wow. That’s a bold statement.

00:00:09:05 - 00:00:43:07 | Vlad

It is it is because they invest so much in the development of talent. And you have a lot of people that are knowledgeable and you can learn from there right. After six years you can do whatever you want. But from learning how to think as an engineer, learning how to work with other engineers, learning how to work with your manager as an engineer. Learning what makes your manager successful as an engineer. Just learning that and learning it in a way that, It’s learning environment like you actually learning something other than you throw it in you have to figure it out.

00:00:44:10 - 00:01:19:21 | Chris

Welcome to Episode 3 of the Engineering Hangout presented by Templarbit. I’m Chris and I’m part of the team here at Templarbit. On today’s episode we have our CEO and Co-founder Bjoern Zinssmeister talking to the V.P. of Infrastructure at Sentry, Vlad Cretu. Sentry is an open source error tracking solution that lets engineers monitor and fix crashes in real time. Here at Templarbit, Sentry is a big part of our workflow so it was great to talk to Vlad about the product as well as his career. Here is episode 3 of the Engineering Hangout featuring our guest Vlad Cretu.

[electronic keyboard music fades out]

00:01:22:22 - 00:01:51:01 | Bjoern

It’s a pleasure to have you on the show. Welcome. Glad I get to record the episode with you actually because we worked together in the past. Tell us a little bit about your new role at Sentry and how the transition has been from you know working for a company that is more than 2000 people to an organization that has a little bit less than 100 employees. Because before since you were actually at Atlassian right?

00:01:51:11 - 00:02:29:08 | Vlad

Yes yes I was. Well the differences are rather big. Right. A large company has it’s own methods that you have to get used to a smaller company is not necessarily crazy but more organic let’s say that. So just trying to understand the existing relationships, respect them and yet make sure that you’re able to do your job is significant. Larger companies have processing frameworks that do that for you. When joining a smaller company you have to be a little bit of a detective and investigate that.

00:02:30:09 - 00:02:39:00 | Bjoern

And tell us a little about Sentry. What is Sentry’s mission? What is your role there and color a bit.

00:02:39:06 - 00:03:16:07 | Vlad

Yeah definitely. So Sentry is an open-source error tracking solution that helps developers monitor and fix crashes in real time. Right. Essentially we provide a number of libraries and in all sorts of languages that folks can instrument in their code and we capture their errors and through that we engage in a workflow that helps developers discover the source of errors so that they can fix it. My role at Sentry is I’m the VP of infrastructure and in that I’m responsible for everything the breaks. [laughs]

00:03:16:13 - 00:03:35:29 | Bjoern

[laughs] That’s great. We are actually a Sentry customer. So we love the tool. We use it actively on all of our apps and I know a lot of people do as well. Like pretty much everybody I recently talked to has either heard of Sentry or is actively using it so you guys are doing something great.

00:03:36:06 - 00:03:55:28 | Vlad

Yes so in that case we are big. we are actually huge in that sense if you look at the developer adoption and the number of companies that are using Sentry it’s way bigger than the company. [laughs] So yes we are a small tiny company. We have about 60 employees. But I believe we have almost 10000 paying customers.

00:03:56:12 - 00:03:57:21 | Bjoern

Wow that’s pretty good.

00:03:57:21 - 00:04:06:06 | Vlad

It’s exciting. It’s exciting. A lot of growth, a lot of opportunity moving forward. I’m super happy to have joined actually.

00:04:06:06 - 00:04:39:03 | Bjoern

Nice nice cool. It seems like you like working for companies that build dev tools because Atlassian you know things like Jira super popular help a lot of development teams move productivity along and Sentry in a similar manner is used by a lot of development teams and organizations that work in tech or build software. What do you find interesting in that space? Like building basically software for other people building software.

00:04:41:14 - 00:05:22:25 | Vlad

I think the mission and the people are different. It’s actually the people. When you look at your co-workers, when your co-workers and the whole mission is towards enabling developers to have a better life, enable developers to have a better relationship with their customers. Right the whole mission of our focus is what makes life interesting again. And in a certain way it even adds some thoughts frameworks to the way conversations are being made the way decisions and fights and conflicts and solutions and problems statements are being made right. They have a different color when you are building a tool for engineers because or for developers.

00:05:23:06 - 00:05:46:15 | Bjoern

Yeah. Would you say that a developer and software engineer makes makes an easy customer. Like when you build software you always think of like the end user. You build a dating product you have to basically picture anybody on earth but when you build for an engineer as an engineer does it make things easy?

00:05:46:15 - 00:05:46:28 | Vlad

Hell no! [laughs]

00:05:46:28 - 00:06:26:12 | Vlad

But what makes it easier is definition of inputs and outputs. Because I think at the end of the day the easy part is that the work flow that we need to enable is easier to define and it’s a little better so you don’t have too much creativity about should we notify you. Hell yeah you should notify me right. So. So that it’s easier to understanding the input like when you define the problem and what you want to get out of it. I think it’s a little bit easier to crystallize that answer but if we’re having a conversation about will this engineer or develop be pleased by the solution. Probably not.

00:06:26:20 - 00:06:42:10 | Bjoern

Yeah. Interesting interesting when you build tools for developers, how do you think design and engineering is maybe different from like other pieces of software?

00:06:42:12 - 00:07:15:15 | Vlad

It’s not different actually because our tools solve problems. Right I think when we are building tools for the developers hopefully the way we get to that problem statement and impact of solving the problem, it’s easier to quantify. But ultimately we solve problems and design helps us think better about how we are solving those problems. We certainly don’t want to be an engineering driven organization or even mindset. The design helps us think better yeah.

00:07:15:15 - 00:07:33:29 | Bjoern

Do you think it’s harder to design software for engineers? I kind of find it harder because the problems are so… because you want the solution to be simple but you’re doing a complex thing like especially in security what we’re up against. I find If I’m designing the right flow and product is kind of challenging.

00:07:34:19 - 00:08:02:07 | Vlad

Yeah well anytime you kind of jump in your own back it’s hard but that’s more of a design. I think it’s more about that as a designer have an easier time designing solutions for engineers right and and maybe that’s the case. Though always I look at how do we define the problems? How do we define the solutions? How do we identify the impact that we want this design or this problem or solution to have right?

00:08:02:12 - 00:08:14:29 | Vlad

And once we start getting clarity around those lines, I don’t think it’s particularly more challenging to design for engineers but I might be the odd one out on this one.

00:08:17:12 - 00:08:36:28 | Bjoern

Let’s talk about Atlassian real quick. So you recently I think it’s a couple of months since you’ve left Atlassian and now you’re at Sentry. We use a lot of Atlassian products as well at Templarbit but I just want to know at Atlassian you were working as the head of engineering for HipChat server. Right.

00:08:37:03 - 00:08:37:12 | Vlad

Yeah.

00:08:37:18 - 00:08:54:28 | Bjoern

And so you were essentially working on a project that gets installed on-prem with customers. When working on software that’s on-prem, what do you think are some of the most challenging things? Is it the update cycle? like getting updates out since it’s on-prem.

00:08:56:04 - 00:09:06:00 | Vlad

Yeah. There are quite a few. So first I’ll say that. don’t think cloud architecture. [laughs] Do not do that. Please don’t do that.

00:09:06:02 - 00:10:12:05 | Vlad

So so oftentimes when we work in the Bay Area we are very cloud driven. AWS everywhere. GCP. Right? And oftentimes when you think about engineers or think about development in the cloud right when building software for that they’re supposed to run behind the firewall specifically on your own premise right. Not private network right. So it’s just your own premise. You cannot afford to think about system design the same way you do in the cloud because you don’t have anyone jumping out of a computer to help you. You don’t have a bank of machines. So it’s a completely different sort of requirements that then I think that mentality is the first one to understand that needs to be changed or adjusted. But you also touched on security updates and maintenance updates and all the other stuff right. And that certainly is a challenge right. Making sure that your customers have the latest security updates in a way that doesn’t actually break your product. [laughs]

00:10:12:15 - 00:10:45:19 | Vlad

Yeah it definitely does. I would say that those are two big significant challenges. Another one I would say is getting customer feedback a lot of behind that firewall customers aren’t necessarily happy with as easy to get feedback as folks that love to work in the cloud right. And like the consumer apps right. So I think that’s really understanding how do you get customer feedback? How do you make sure you are building the right software for the customers? It is challenging in and of itself.

00:10:45:19 - 00:10:56:16 | Bjoern

Why do you think they’re so coy in giving feedback? Like is it just that those on-prem customers are just a little bit more old school? Where do you think that comes from?

00:10:56:23 - 00:11:23:02 | Vlad

Well because you have two customers on-prem. You have the administrators that actually administer your products and then you also have your end users that are using your products right. And and you also need to recognize that they are enterprise people right. So they have their jobs to do usually right. And maybe their relationship with the software is very different than the relationship with software that when you have when you engage with cloud products.

00:11:23:15 - 00:11:43:20 | Vlad

And I’m only speaking particularly not for like a generic product but something that is niche right. Like utility that I just want to get this done and maybe I’m going to tell it how I like it but liking it is not really important to me I just need to get the stuff done right. So I think there’s even a different relationship with how we identify value.

00:11:44:03 - 00:11:57:25 | Bjoern

How important do you think is the on-prem universe? Do you think it’s eventually dying completely? Do you think it’s going to keep being very relevant? What are your thoughts.

00:11:57:25 - 00:12:16:19 | Bjoern

I know it’s hard to predict the future but like you’ve had a pretty recent exposure into shipping on-prem software and I’m always curious like how this space evolved. Is multi-cloud going to be a thing for a long time? Is everybody eventually going to move into the cloud completely?

00:12:17:03 - 00:12:25:18 | Vlad

Yes so there’s no getting away from the cloud because the folks managing the cloud are likely better than your I.T. people or my I.T. people right.

00:12:25:18 - 00:13:12:23 | Vlad

I think just if you think in terms of the skill set to keep everything up to date even from a security perspective and maintenance that the folks managing the cloudss are way better than what I can do or what the people I hired can do right. So that alone is why the shift to cloud is happening. The question is like is it going to take 100 percent and from my perspective it’s going to take ninety nine point ninety nine percent because there will always be companies that cannot do that. For example, I’ll be very surprised if Intel will move anything to the cloud right. But only because they probably have the resources to run it inside right. So outside of those folks that are the size of intel like people right most of everything will be in the cloud. And it should actually.

00:13:13:15 - 00:13:32:02 | Bjoern

Cool. A few years ago, you worked at DocuSign and McAfee and Microsoft I’m bringing up this question because we just talked about Intel and when you worked at Docusign and McAfee, Microsoft those are also companies that are very very enterprise, they have enterprise in their DNA.

00:13:32:11 - 00:13:58:10 | Bjoern

And what are some of the interesting aspects of building products or working in an organization that is so like very enterprise compared to doing dev tools or consumer apps? What are some of, I know it’s been a couple of years yet worked for them, but like what was that time like? It was like early days a second wave in the valley but these are really large companies doing large things.

00:13:58:14 - 00:14:10:06 | Vlad

Yeah yeah and you know they have a lot of things in their DNA not just enterprise. I loved my time at Microsoft. I would be a completely different person without the Microsoft experience.

00:14:10:23 - 00:14:15:27 | Vlad

I would even go out to say that every college hire should work at Microsoft for six years.

00:14:16:01 - 00:14:16:04 | Bjoern

Wow. That’s a bold statement.

00:14:16:21 - 00:15:05:06 | Vlad

It is it is because they invest so much in the development of talent and you have a lot of people out there that are knowledgeable and you can learn from there right. After your six years you can do whatever you want but by learning how to think as an engineer, learning how to work with other engineers, learning how to work with your manager as an engineer, learning what makes your manager successful as an engineer. Just learning that and learning it in a way that it’s learning environment like you actually learning something other than you throw it in and you have to figure it out. It’s invaluable, it’s invaluable. Some people maybe have that experience to other means. But I’d say most folks would benefit from getting thrown in that environment and just seeing how things are being done.

00:15:05:08 - 00:15:25:01 | Bjoern

I’m actually also a huge fan of and I give this advice a lot too young or recent grads… Well sometimes I tell them to go join an early stage startup because I run one [laughs] Actually oftentimes it’s the correct move in my opinion to go to a very established corporation, old-school dot com, whatever you want to call it.

00:15:25:24 - 00:15:57:10 | Bjoern

And for the same reason it’s a really great environment to understand what is a product manager? What does that guy do? How does QA work? All those things that are hard to learn in a smaller startup or even like an A round, B round startup where that stuff isn’t yet in place or if it is in place it’s barely in place and spending like, maybe not six years I would say maybe more like two or three, and seeing that and experiencing that makes you a much more effective engineer at a startup later on.

00:15:57:16 - 00:15:59:09 | Vlad

Yeah definitely no argument there.

00:15:59:10 - 00:16:16:22 | Bjoern

Yeah. Cool. Let’s dive into engineering management I want to ask you about that since we’re talking a little bit about it already. Over the years I’ve noticed that a lot of startups when they hit sort of the point of 20 plus engineers that sort of where it starts.

00:16:17:06 - 00:16:28:21 | Bjoern

They start to have challenges in that part of their business. Doing engineering management, getting it done right. You will see a lot of maybe retention problems especially in the Bay Area.

00:16:28:22 - 00:16:40:12 | Bjoern

What do you think are some methods, some learnings that you can share with people that they can implement? That helps them run things a little bit more smoothly.

00:16:41:12 - 00:16:53:22 | Vlad

Wow. [laughs] There are tons of books are written on that. But I’d say the first thing you need to really understand and everybody needs to recognize is what is the purpose of engineering management?

00:17:02:04 - 00:17:03:13 | Bjoern

What are some of the answers you’re heard?

00:17:04:22 - 00:17:06:04 | Vlad

I just want people to…

00:17:07:15 - 00:17:08:12 | Bjoern

Be more productive Yeah…

00:17:08:16 - 00:17:12:12 | Vlad

I want shit done now, right and I need somebody to drive that stick.

00:17:13:06 - 00:17:32:28 | Vlad

And not to necessarily emphasize on the negative connotations of that but there is a certain sense of urgency captured in that message as well and very much relevant for startups as well. Right. But overall there needs to be a very explicit, almost declarative, way of what we expect out of this engineering management function.

00:17:32:28 - 00:18:01:28 | Vlad

Right. And that needs to be recognized and understood by everybody. Right. That is… when that is missing that everybody has their own private interpretations or understandings of what the hell that means. Right. And that is when you start having conflict. Right. So I’ll say that is that the number one thing to do. The second one from a tactical perspective in the sense of what are the things you can do to be more successful in that right.

00:18:03:05 - 00:18:26:10 | Vlad

I would say that things that generally get overlooked and even by myself if I’m too much in a hurry is write things down. Like always write things down don’t assume closure or don’t assume that everybody understands or everybody walks away from a meeting with the same things you walk away from the meeting.

00:18:26:10 - 00:18:46:27 | Vlad

Right. I think that there’s just this very basic hygiene like behaviors. Right. Do what you can so that folks really understand what’s the outcome of a meeting? What is the expectation of a work teakettle or of a problem right.

00:18:46:27 - 00:18:58:14 | Vlad

Like how do you talk about the results that a release is going to have right. So the more you talk, the more you write those things down, the less issues you have down the road.

00:18:58:27 - 00:19:21:16 | Bjoern

So it seems like you could sum up that in a way communication or certain type or specific communication and establishing a baseline of like what you can expect from your team. Those are some of the most impactful things you can do as an engineer org.

00:19:21:18 - 00:19:36:02 | Vlad

That’s one of them and if I wasn’t as verbose I’d probably have summarized as nicely as you have right. Just making sure there’s no confusion, there is a record. This is a trick and this is what I do right when I want to communicate.

00:19:36:02 - 00:19:43:15 | Vlad

I’ve learned that I should communicate via e-mail because I want people to have something to reflect on.

00:19:43:21 - 00:19:57:15 | Vlad

Yeah like if I go into a room and I communicate in the room within five seconds after I uttered a sentence your mind is somewhat outside that. I don’t control that environment and I don’t know how I’m going to get quality from our interactions right.

00:19:57:20 - 00:20:12:25 | Vlad

So if I have something that is a little more meaningful to say I usually always do document so that I can reflect on it later or someone can reflect on it so we can interact with those ideas in a little bit more purposeful and maybe a little bit of a mature way.

00:20:12:26 - 00:20:14:12 | Bjoern

Yeah it’s a little more meaning.

00:20:14:14 - 00:20:15:16 | Vlad

Yeah yeah definitely.

00:20:15:24 - 00:20:39:20 | Bjoern

Interesting. That’s cool. Let’s switch gears a little bit and talk about San Francisco. So you’ve lived in San Francisco for a number of years, probably longer than I have. And imagine you showed up here tomorrow because it’s a very different place than it was even five years ago.

00:20:39:28 - 00:20:56:26 | Bjoern

If you showed up here tomorrow, maybe you’re a younger engineer, maybe you’re not a younger engineer whatever, but you coming from the outside and you are trying to either make a name for yourself or maybe you want to make a little money doing a startup dream. What would be one of the first things you would do?

00:20:57:15 - 00:21:07:25 | Vlad

Live in a hacker house. Don’t care which bed I wake up in. Meet smart people and try to figure out how I can help them. I think I think that’s the three things I would do.

00:21:08:09 - 00:21:13:05 | Bjoern

So basically get super connected as quick as possible.

00:21:13:23 - 00:21:48:03 | Vlad

Super connected with people that I want to build alliances with and so that you have some responsibility to detect like hey those are interesting people that I want to work with. Right. So connected. That’s right. Then you have to manage your cost. So that’s why I say sleep in a hacker house don’t get anything else right. Don’t care where you wake up and often times you will have a good time. But just reduce your cost of living. Right. And meet really smart people. I mean there’s still a high concentration of folks up that want to change the world and you should see them.

00:21:48:05 - 00:22:14:18 | Bjoern

Yeah. So you would prioritize making connections with humans over let’s say another strategy which could be, alright there is a ton of jobs available, a ton of opportunities to get a really great stock options in different verticals. So you wouldn’t necessarily show up and look at all the available jobs and stocks and chase that you would rather chase connections… humans?

00:22:14:27 - 00:22:49:00 | Vlad

Not that you don’t have to look. If you need to look if you have a huge college loan that you have to go for it I mean sure get a job right. Yeah but if you can afford to not work eight hours a day in an institution right. If you can just afford to meet people if you can afford to be curious you can afford to build relationships with people that are as curious as you are and as driven or might be even more driven than you are. Then optimize for that you can always make money. I mean you are here in San Francisco. You can always make money.

00:22:49:21 - 00:23:08:26 | Bjoern

Awesome. S.F. San Francisco or even the Bay Area as a whole, I think one of the best things about living here, which I haven’t really thought about until recently and I don’t take advantage of this nearly enough, There’s a ton of conferences here.

00:23:09:06 - 00:23:21:13 | Bjoern

And if you live here it’s actually relatively cheap you know to get in there. Mean you only get an expo hall pass or you know a friend of a friend and you get a ticket to Google IO for free or whatever.

00:23:22:07 - 00:23:32:15 | Bjoern

What are some of the conferences that you think are cool? What do you go to every year? Do you have a set conference? What do you think about the conference scene?

00:23:32:28 - 00:24:37:06 | Vlad

Yes I’m a cheap date. I like meet ups. [laughs] Get me a beer and you have your way. But look I llike meetups. I think meetups it’s like it’s a catalys,t it’s a secret that we have. It’s a culture we have to protect. At meetups is when you meet people that usually tend to have an influence on you or if you make the effort they’ll have an influence of you like. Conferences are great to people watch and to get a marketing talk right. But there’s no substitute for what meetups can do in terms of opening up opportunities to just learn ideas. Learn how to be curious… meet people. Usually what I like about meetups the most is it’s like… Do you notice when you go to a meetup and you see everybody kind of enthusiatic and smiling and all the smiles and looking they just pushing to you like you always feel them physically the presence. It’s romanticizing what’s in the air is it’s quite exciting. So anyway that’s what I think about it

00:24:39:15 - 00:24:46:17 | Bjoern

That’s cool. So you prefer meetups over conferences. Yeah I think I do too but I haven’t gone to one in a while. What’s the last one you went to?

00:24:47:09 - 00:24:52:03 | Vlad

Two weeks ago at Lyft. It was on data analytics.

00:24:52:03 - 00:24:52:15 | Bjoern

Nice.

00:24:52:16 - 00:24:55:23 | Vlad

And their pipeline and SQL.

00:24:56:01 - 00:25:33:15 | Bjoern

Nice. That’s interesting. Yeah I have gone to a meetup in SF, not even this year I think. I’ve been doing the conference scene a little bit more and I’m getting more appreciation because I used to do a lot of meetups and now you know because of the industry that I’m in I have to do a bit more conferences. I’m starting to appreciate that I; actually feel like guilty that I haven’t gone as many conferences in the past living here and having access to it you know, not having to worry about a thousand bucks a night motel rooms. So I’m a little bit more on that track but now I feel like I need to go to a meetup again.

00:25:33:18 - 00:25:52:12 | Vlad

Yeah but even if you don’t… I used to have a habit when I go to meetups even if I’m not interested in technology because I just want to listen. Yeah right. I’ve fallen off the wagon but reengaging in that it’s… it’s… energizing?

[electronic keyboard music fades in]

00:25:52:29 - 00:26:01:27 | Bjoern

Cool. Yeah that’s the end of the episode. So thank you so much for being on the show. I really enjoyed the conversation. Thank you so much.

00:26:02:00 - 00:26:06:04 | Vlad

Hey man, It’s been a pleasure. Thank you.

00:26:06:06 - 00:26:26:07 | Chris

That’s it for episode 3. I’d like to think Vlad Cretu for being our guest today. You can find out more about Vlad and Sentry at Sentry.io. You can subscribe to the Engineering Hangout on Breaker, Apple podcast or anywhere else podcasts are found. On behalf of the team here at Templarbit, thanks for listening.

[electronic keyboard music fades out]