Getting a remote job at Auttomatic with Leif Singer
We also talk about:
- why he switched from academia to industry,
- how he got hired without knowing the company’s tech stack,
- why he worked for several months at Automattic before being hired,
- how the Automattic creed, influences, and guides company values and mindset,
- Automattic’s software engineering practices,
- and how they fight technical debt during quarterly scheduled clean-up weeks.
Leif Singer, is an engineer at the fully remote company Automattic. Automattic is the company behind WordPress, a famous and powerful content management system with an estimated 75 Million user base.
Read the whole episode "Getting a remote job at Auttomatic with Leif Singer" (Transcript)
Michaela: [00:00:00] Hello, and welcome to the Software Engineering Unlocked podcast, where I talk to experienced developers from different companies about how to develop software. I'm your host, Dr. McKayla, and today I have the pleasure to be talking to Dr. Leif Singer, who is a software engineer at Automattic. I'm really excited to have Leif here today to hear how his life is at a fully remote company. Leif is like me, passionated about improving software engineering, processes and tools.
His specific interest is understanding how lightweight processes can help to collaborate better. He has a PhD in software engineering, but is also a science fiction writer and a musician. So I'm thrilled that Leif is here today. Leif welcome. How are you?
Leif: [00:00:44] Good morning, Mikayla. Thanks for having me. I'm good, thanks.
Michaela: [00:00:47] I'm glad that you're here. So Leif let's start at the beginning. We actually met when we both were working and living in well, beautiful Victoria in Canada, precisely at the university of Victoria, where you worked as a postdoctoral researcher. Now we live in Germany and work as a engineers of engineer at Automattic. That's quite some change. So yeah. Why did you change from academia to industry and how did you start working at Automattic?
Leif: [00:01:13] Wow. Okay. You're you're starting with the tough questions, interesting. So you know that, but not everyone might know that, a postdoc position is a research position at a university that is supposed to prepare you for the job hunt for a professorship, right?
So, I did the post doc because I thought I wanted to become a professor. But when I worked as a postdoc with two supervisors who were themselves already professors. I realized that academia was after all, not for me. The amount of time that I saw my superiors, and my colleagues too, put in, far exceeded what I thought was reasonable.
They loved their jobs. So they worked on weekends and things like that, right. And I, as a father of a young family, I didn't see myself doing that regularly in the near future. And that was first thing that crossed my mind. And then a few other things came together with us living in a, in a different country, far away from home with no support network of our own with a small son. And in the end, we decided to cut the Postdoc short at one year and moved back home.
Michaela: [00:02:35] Okay. So it was actually supposed to go longer?
Leif: [00:02:38] Yes, it was supposed to take two years, but I talked to my supervisor and she was all right with me cutting it short for one year.
Michaela: [00:02:48] One of the things you also addressed a little bit was that you were living in a different country. And I also traveled and lived in different countries over the time, but I always had the feeling that I couldn't choose it. But at the point that you want to become a professor, it's somehow out of your hands a little bit. It's not like you can say, well, I want to become a professor in Germany. That's really a very difficult task.
Leif: [00:03:10] Right. You can't be choosy at that point. You have to just move where the jobs are, and that was definitely part of our decision.
Michaela: [00:03:18] I totally understand that. So then you made this decision and what happened next? What was the next thing that you went on to do after your postdoc?
Leif: [00:03:27] I didn't have a plan. So I looked for a job and I found a remote job at a small company called IDoneThis. And it's a small app. When I joined, it was me and four others who worked on that app, it was really small, based mostly in New York. I worked there for, I suppose, one and a half or two years almost when the founders at some point, decided that they want to sell the company. And that's when I started looking for a new job because the buyers didn't want to buy the team, only the product and the customer base.
Michaela: [00:04:09] That's a bummer, probably, if you put part of your heart in there.
Leif: [00:04:12] Yes. Yes, definitely. It was a bummer, but in the end it turned out that Automattic were some of our customers. So I got an intro from our CEO to one of Automattic's, basically the person who managed the Automattic account. And that's how I even thought about applying to Automattic and how I found out about this company. So in the end, it turned out to be a nice circumstance.
Michaela: [00:04:42] When you were interviewing, then, for Automattic, did you have the skillset that was needed for that position or did you have to adjust there?
Leif: [00:04:50] So Automattic is the company behind Wordpress.com. And as such, externally I would have assumed that you need to know PHP and WordPress development. I had not developed a single WordPress plugin or theme before applying. I did have a few years of experience with PHP when I was doing my bachelor's at University, I basically worked on the side as a PHP freelancer, but when I went through my trial, at Automattic, you go through a trial over the course of one or two months where you work on basically a real project. I worked on a React app that was not released at that time. It is now it has now been released as the front end or, well, the administrative backend for Wordpress.com. So, pretty much a replacement for the WP admin, part of things written in React JS with Redux, server side rendered with Node, lots of buzzwords in there. And I had never heard of anything from those technologies before I started working on that, but that was also part of the process. I believe of the trial process to see how I learn, how I adjust to new technologies and things like that. What kind of questions I ask.
Michaela: [00:06:11] Okay that's really interesting. So you started with this, maybe let us look at a little bit at this trial, because I think it's very specific to Automattic how they hire, right? But once you explained that, maybe look at that you actually started that without prior knowledge of the technology. I think that's really mind blowing somehow. I really liked that concept, but yeah, it can be frightening. I can imagine.
Leif: [00:06:32] It really was. And actually, I don't think it's that common so I can see someone of what current trials are working on and it's much more matched to their skillset. So I suppose I got lucky in an earlier time of the company where that was possible right now. I think we are trying to standardize a bit because we really need to scale up hiring. And that is much easier when you have one approach that works.
Michaela: [00:07:04] So how does this such a trial look like at Automattic right now, if I want to have a job at Automattic, what do I have to do?
Leif: [00:07:09] Right. So at first you send an email, you, or it might have changed, actually, maybe you fill out a form, we're starting to use the tool for managing those applications as well. Anyway, you write in with your CV and a cover letter. And at some point, hopefully soon, you hear back with hopefully a yes, we'd be interested.
And then you do an interview over text. So. I believe we use Slack nowadays, you get invited out. Yes. It's a text chat. And you basically talk for about an hour where you're being asked about your career, about some technical things. So it's basically a screener where we try to figure out whether it would be a good match thinking wise technology wise, things like that.
If that chat works out well. You get invited to do a small project that should take, I don't know, the actual times right now, but I think it's about four or five hours that you need to put in over a week where you get a small piece of code that you are supposed to improve with specific instructions. "Hey, is this code secure", "What if we add this and that feature, could you do that?". Things like that. And that's, I think we call that a code test, but if that works out also, then you get to do a trial, which can take anywhere from one month to two months. Some I've also taken longer. You get assigned a real project that we really want to use and sometimes you're part of an actual team.
You get to know a few of your possible future colleagues. And we basically look at how you write code, how you communicate, how you react in the face of challenges, how you scope things, things like that. We feel it's a very wholesome picture that began of the applicant, it is also paid at a flat rate of $25 per hour for everyone. It's no indication of, your future pay. But it's the same pay for every role at, during that trial. The trial should be doable, well, putting in five to 20 hours per week, I suppose, I think I put in about 10 hours per week, which can be a challenge if you're working and maybe have children, we try to accommodate as many people as possible.
Michaela: [00:09:40] Sounds interesting. So I'm working alongside the team that I will probably work with afterwards?
Leif: [00:09:46] Well, maybe, maybe. That's not definitely, but you will work with other Automaticians, or Automattic employees.
Michaela: [00:09:57] And then I will do daily tasks that are needed for a position actually at that team. Right. So I really do what they would expect me to do?
Leif: [00:10:05] Yes. And you will write code that if it works out well, will probably be used in production.
Michaela: [00:10:13] I see. Well, I actually used throw you a little bit in the deep end, right. With the whole hiring processes and things like that. And that's actually not your role, but I think it's always interesting to see a different ways of entering a company.
Leif: [00:10:26] Yeah. I agree. And it's one of the, and it's one of the interesting, more interesting, parts of the company where it can see how it scales.
Michaela: [00:10:34] Yeah. Because it's very different, right? But yeah, coming back to Software Engineering, at Automattic. So you talked about React and Nde.js and things like that. What is the tech stack? So if I want to apply there, or what do you have to bring to the table to be able to work there.
Right. So, for example, one project is based on GitHub and it's written in React JS. It's a huge app, um, uses Redox as a store and we have proper continuous integration there. We have unit tests and integration tests and end to end tests and whatnot. And then another project, uses, Subversion, and PHP and WordPress, and that's pretty much it,that's all the technologies you have available to you. So it, it can differ or it can vary a lot. So it helps if you come to it with a certain generalist attitude, I think. And if you can learn, then I think you already have the most important skill.
Michaela: [00:12:22] So what you're saying is also that you're maybe moving within the company. So if I let's say I started, on the Git project that you just mentioned, and then it could be that I'm actually moved or I want to move. I don't know. What exactly happens to move to a different project.
Leif: [00:12:35] So I started out working on the React app, which was very interesting with all those new technologies and everything right. Using Git is always very pleasurable experience, and that is without sarcasm. But at some point I wanted a change of scenery. I wanted to learn other things that I had not learned about before. Um, so right now, how that works usually you just send a private message to Matt and tell him. I am looking to change teams. I'm interested in these things. Can you take a look? and then after a couple of days, he gets back to you and tells you that sure. We might have a position open and this team, what do you think? And that's how he started the process of switching a team internally. So now I work on the marketing team because I thought that being a developer and knowing something about marketing, that sounds like I'm like a super power combination, you know? Like, knowing how to market a product that you can also build. So right now I'm really in the depth of that, specializing a bit now. I'm a project lead for affiliates and partnerships where we basically give people a commission if they refer customers to us. So that's what I've been doing in. the past months, I guess.
Michaela: [00:14:03] I really find it an interesting approach that you send an email to all the CEOs
Leif: [00:14:08] It's not an email, we don't use email pretty much. So it's a Slack DM / Slack message.
Michaela: [00:14:15] And then he probably asks around for you or talks with leadership somewhere else. And you figure out where and how that could be.
Leif: [00:14:23] Exactly. And he also talks to the hiring team because those are the people who naturally need to have good understanding of the open roles that we need to fill.
Michaela: [00:14:32] I guess that this process looks different from other companies because you are a fully remote company, right? So you can not bump into somebody at the water cooler and tell them, "well, you know, actually I'm a little bit interested in changing teams", or something.
Leif: [00:14:44] Well, you can do that too. So when you've worked on a few teams, you will know a few people around the company who are maybe not in your current teams. So you can also use those connections to figure out where there could be a good match for a new team.
Michaela: [00:15:03] And, but how do you keep contact with those people? Do you use Slack for that or do you sometimes hang out somewhere in the cyberspace togther?
Leif: [00:15:14] So we use Slack for the more synchronous communication needs. Most teams have weekly calls. Most people will have a weekly one on one call with our lead. And those would usually be a video calls. We do have a, a company wide town hall once a month, where Matt basically hosts a huge call and talks about what's going on in the company that month. And everyone can join from the 900 plus people. Everyone can join that one call and ask questions through Slack. It would be, pretty messy, I suppose if everyone was talking and then we have WordPress, of course we use a specific or a special theme called 'P2' that we use for project and team blogs. So every team and every project or larger projects will have their own blog, where we can post updates, questions, specifications, documentation, communications with outside vendors and stuff like that. And that is pretty much all we use for communicating with each other. We do not use email at all.
Michaela: [00:16:25] Wow. That's so different to my experience at Microsoft, where we are fueled by emails, right?
Leif: [00:16:32] So another big part of how we get to know each other and how we build up trust is meeting in person from time to time. So every team meets about two times per year. And once a year, the whole company meets in one single place where we have one week of meeting each other, working on projects, learning things together and stuff like that.
Michaela: [00:16:53] Yeah, that sounds really good because otherwise you will have never seen them in person, ever.
Leif: [00:16:58] Yeah, and it's also to be honest, it's, um, a big perk of the company too, that you get to travel like two or three times to interesting places.
Michaela: [00:17:06] Yeah. I always see your Instagram pictures, then, where you say like, what was it New Orleans or something was the last time. No?
Leif: [00:17:12] Yeah, yeah.
Michaela: [00:17:13] Yeah New Orleans, right? Yeah, wow. It's amazing. I would like to be there as well.Yeah. So when you look at it, so your company, is there something there's specific going on for the processes and tools that you use for an enormous software development? Well, really writing the code, building the code, deploying the code, testing, something like that. Is this something that you think is different from other companies and how does it look like in general?
Leif: [00:17:39] I don't know about other companies that much. Um, but I can say that because we have a bunch of different products and they are all technologically, somewhat different, the approaches we used to: So for continuous integration and for version control, for example, they can differ like by decades. So WordPress is on some version, right? And we have a WordPress repository and we need to, to merge with the WordPress core project from time to time, or we have to merge the open source project into our repo from time to time, right? And it also has our own modifications to it. So that project is on Subversion and there is no GitHub for version, right? The best or the next best thing that comes close to that would be Fabricator, which is, kind of like an interface on top of Subversion repositories, where you can also do things like pull requests and get reviews on those and stuff like that. And that system itself is also, again, we don't use Circle CI or Travis CI or things like that on that repo because those systems assume a Git based repository. So we use something that's called Team City on that. So basically the parts are all the same, theoretically. But in their manifestation, they differ. Like they, they will use different technologies. So we do use version control. We do use code reviews. We do use continuous integration. We rewrite tests, all those things, but how we do that and how hard it is also will differ from project to project.
Michaela: [00:19:19] Yeah, that's really interesting. I mean, how big is Automattic? Can you remind me again?
Leif: [00:19:26] The company? That's just 900 people,
Michaela: [00:19:28] 900 hundred people, Okay. And how many engineers do you know?
Leif: [00:19:31] It's a rough estimate, I think like, I don't know if I recall exactly, but maybe about 400.
Michaela: [00:19:37] Okay. So for 400 people, I think it's quite diverse, right? I mean, we are dealing with all those different technologies. You're dealing with all those different tools. So also the knowledge that you have to have to know all of those really in and out is, Quite scattered, right. Is there some future goal of consolidating those different technologies? So like for example, porting the WordPress code base to Git or is that too huge a thing to do.
Leif: [00:20:06] I think that's politically very heavy subject, because we depend on the WordPress open source project. Right. And it's an open source project. We don't get to dictate what they use, right. So we have a whole team who basically gets paid to work on the open source project, but they are trying really hard not to let company decisions or influences, let that influence their open source work. So we really. Depend on the open source project for that. So if the open source project moves to Git, we can move over too.
Michaela: [00:20:43] I see. Okay. Yeah. I never thought about that. I mean, that's really an interesting aspect as well of working at Automattic that you have like this, I call it now tension a little bit tension between open source and the commercial interests of a, of a company, right?
Leif: [00:20:57] Yeah.
Michaela: [00:20:58] What are the benefits that the company gets out of open sourcing parts of their software. Why do you do that?
Leif: [00:21:06] We are an open source company. The company is founded on the idea of open source basically, right? So we embrace it and I'm not sure if it's really about the concrete advantages of open source, but I think culturally, the spirit of openness of transparency, that's helpful in other areas too.
For example, if all the communication is transparent, And if decisions are transparent, that creates a vastly different company culture, and that might not directly affect code or things like that. But it will affect many other things.
Michaela: [00:21:39] Yeah and actually, I have to say that when I asked you to be on my podcast, it was quite a quick decision, right? It's just send an email or, not an email, a message right? To one of your colleagues and they approved it. And I know that for several other interview participants, it's like weeks for them to get well, allowance to even talk. And some of them are just purely denied. It's like, no, you cannot talk on a podcast.
Leif: [00:22:06] So we have a Slack channel called PR for public relations where people talk about things like that. And I just asked the question there and by coincidence, the colleague who replied to my question was Matt. Okay. So I got official okay from the CEO.
Michaela: [00:22:26] Oh yeah now I feel really at ease. Now I can ask all these tricky questions.
Leif: [00:22:32] Yeah, exactly.
Michaela: [00:22:35] Yeah. Very cool. Yeah, that's really something. I mean, it's a minor peek into your company, but for me it was very. Well, relieving also. I asked you to ask for permission, and you asked for permission and within a few minutes, you're coming back and saying, yes, I can talk on your podcast. And that's so nice, right?
Leif: [00:22:52] I mean, we do have some, some back and forth too. We don't always have transparency, even if we want it, but the default is transparency, openness, low bureaucracy. So, we will never reach the ideal, but we're always trying to approximate it. Right? So one guideline that Matt keeps on mentioning is to quickly make decisions that are easily reversible, you know. Like don't, don't fret too much over making a decision that you can reverse within the matter of minutes anyway.
Michaela: [00:23:27] Yeah, that's really smart. So there's a lot about culture, right? How the culture of a company works and well, how long have you been now at Automattic?
Leif: [00:23:38] Ooh, I am now, I joined three and a half years ago.
Michaela: [00:23:42] Okay. So do you feel that the culture, I mean, the company evolved and grew quite substantially?
Leif: [00:23:49] Yes.
Michaela: [00:23:51] Did the culture also change much during that time or did it stay the same?
Leif: [00:23:57] Both. I think so. I think when I joined who we were about 300 people now we're over 900. So it has tripled in size within three years. It would be pretty weird if that would not create friction, I suppose. Especially with people coming in who might have worked in other contexts who might be further removed from the open source project, for example, they will carry a different cultural inventory with them. I believe. So it's a constant balance that we're trying to strike between letting more diverse perspectives in, but staying true to the core of the company values. One thing that helps with that is that we have a creed. The Automattic creed. So if you go to Automattic.com/creed, "C R E E D", you will see a short text, which is something, "the like the vision for the culture of the company in a nutshell". And I'm going to read it to you. And the Automattic creed says,
"I will never stop learning. I won't just work on things that are assigned to me. I know there's no such thing as a status quo. I will build our business sustainably through passionate and loyal customers. I will never pass up opportunity to help a colleague, and I will remember the days before I knew everything. I am more motivated by impact than money, and I know that open source is one of the most powerful ideas of our generation. I will communicate as much as possible because it's the oxygen of a distributed company. I'm in a marathon, not a sprint, and no matter how far away the goal is, the only way to get there is by putting one foot in front of another every day. Given time, there is no problem that's insurmountable"
and that's the Automattic creed. And that is basically the, the gospel that we try to reference when we need to make decisions that we're not sure whether they align with company goals, company, culture, or values and things like that. And usually it at least gives us a rough direction to turn to.
Michaela: [00:26:12] Yeah, I liked it. I liked it a lot. It sounds like a person spent a lot of time really thinking about the values of the company. Yeah, that's nice. So I wanted to switch gears a little bit now and talk with you about the software engineering processes. So do you have similar guidelines, like the Automattic Creed for your development processes, and how do you do things? Like, for example, how do you test your system?
Leif: [00:26:38] So non end to end tests, all the tests that don't need a browser to run, I would say the developers to write themselves as part of their ongoing work, but for end to end tests where we use like a handle as browser to actually use the site and buy things and sign up for that, we do have a specific team who works on that, just because it's a very specific niche technology and you have to develop a very deep expertise to be able to automate all that and to know how to handle issues. For example, if you try to buy something and the test fails, why is that? Is that because maybe the payments provider was down or is it a problem in your code? And if you have many dependencies to other services, then it can become pretty hairy pretty quickly. So it's nice to have a specific team for that.
Michaela: [00:27:36] Okay. And do you have like something like code coverage? Do you track that? Is this something that's part of your daily basis or not?
Leif: [00:27:42] No, I don't think we do that. I think we did discuss measuring something like code coverage at some point, but then I think we didn't want to focus too much on a single metric. So we did experiment around that a bit. And I think what we ended up with is a integration on GitHub that posts a message to your PR. If the code that you add has code coverage lower than something, something like that. Something not too obnoxious, something kind of subtle.
Michaela: [00:28:20] And now that you said, well, the whole landscape is very diverse technology wise, but also tool-wise. Is, is that also something that's different for each of the teams and projects that you work on? Or is there some standard policy around testing, code reviews, things like that?
Leif: [00:28:37] We don't have a standard policy written out anywhere, I think. We do have guidelines in, so we do have a company handbook, So to say. It's called the field guide. It's just another blog. Everyone can edit it and you can find out pretty much anything you need to know about working here by looking into it. And it does have a few guidelines for how to, how to test the projects, um, how much effort to put into it. So you don't get derailed, but it's no, there's no hard written rules, I would say. So usually you would. Look at the product that you're working on and figure out what the testing culture is on that product. And then basically blend in, be normal right.
Michaela: [00:29:24] Okay. But it can differ from project to project.
Leif: [00:29:28] Yeah. Especially because it's just different, very different technologies. Right.
Michaela: [00:29:33] Okay. So it's sort of technology driven, alignments.
Leif: [00:29:38] Technology influenced, I would say, yeah. And also, um, I mean, yeah, different products are obviously of different importance, right? So that will play into that as well. I would say so Wordpress.com has, I don't, I don't know if I can say the number, but it has lots of users and it would be a really bad if it would break. So that would have different standards than say an internal app that we're developing for just for comfort reasons.
Michaela: [00:30:08] Yeah, that makes sense. And what about code reviews? So, one of my topics that I worked a lot on was code reviews and how they are used to improve the code quality, but also to share knowledge amongst developers is, is contribute something that you do at Automattics and how are people perceiving it?
Leif: [00:30:27] I think it's changed over time. So when I joined, there was still a few people who would just cowboy code and commit, but that is getting fewer and fewer. It's been. Probably two years or something like that since I've seen anyone commit without a review. So every pull request or whatever you may call it, every change will have been looked at by at least one other colleague. If it's very sensitive stuff, like for example, payments code, then we would have two or three other colleagues look at it, depending on at what level we would feel comfortable. Right. We'd also reach out to experts on the subject matter if we are not experts ourselves. So that's yeah, I think that's really nice how it works. I feel very happy with having everything reviewed.
Michaela: [00:31:16] Mhm, this gives you a good feeling about your own code and about the code from others?
Leif: [00:31:20] Yeah, generally I think it makes for a good environment where we all know about our own fallibility. Like, we all know that "this is an obvious fix" were famous last words. Right? Like it's very hard if you work on something to really, really, really be sure that this doesn't break anything because you don't know everything. Right. And again, going back to The Creed, "I remember the days before I knew everything". So this sort of humility is kind of built in into the company culture. So most people are very thankful for, for others, um, to look at their code.
Michaela: [00:31:57] Yeah, that sounds great. Especially for code reviews, one of the pitfalls or problems with it is that you wait very long for other people to look at your code. And I can imagine that the distributed company, while remote and distributed company with different time zones, that can be painful. How do you mitigate that?
Leif: [00:32:17] So many teams use, um, Trello. To basically move cards through backlog, this week, in progress, in review, and done. Something like that. And those changes will be posted to the team's Slack channels. So if I see a colleague move something into a review column on Slack, then I'll know this is ready for review.
I will usually not get to it immediately. But if it's, it has some time pressure behind it, then the colleague will reach out to me and ask me, Hey, do you have a minute today to take a look at this? I really like to deploy this today because that would really benefit our customers in X and Y. But there are also of course, reviews that are neither important nor urgent, but we still are one to review that code.
So we do have a, it's a Slack app. We call it the stick and the stick. So I think it comes from being beaten with a stick. So it's like, it's a nudge, you know, it nudges you, to review. So it has like a config file where it can set up your teams and those teams Trello boards. So it continuously, or once a day, it looks at your teams Trello board looks at all the cards that are in the in review column, and then randomly pings people on Slack "Hey, Leif, what about reviewing this one today?". So you get actual pings on Slack by the stick and depending on the person, I think very few people ignore it because a ping on Slack is kind of like, it's a thing you kind of should be doing okay. Yeah. And that really helps keep that "needs review" column clean.
Michaela: [00:34:12] Okay. So one thing that I wanted to ask you as well is, well, if you have this, do you have this review staff, but then how far are actual developers away from production. What I mean by that is, well, imagine you're making a new feature or back fix or whatever, what steps are in between you saying, well, my code is done and having it in production, how much steps are in between which steps and maybe a roughly estimate of a time, if you can give it.
Leif: [00:34:42] So again, the specifics will depend on the project that you're working on because of the different technologies. But roughly if you have something that has been reviewed. So you're confident that a sufficient number of reviewers have told you that this can be deployed, then you merge it and you deploy it. We have a command to deploy the latest master branch, basically, to production. And that's it.
Michaela: [00:35:10] Okay. So you're quite quickly there and that's then really in production, is there some staging happening before, or?
Leif: [00:35:17] We do have a staging for at least one of the projects where we can try things out with the proper host names and everything, and with actual production data, but that's like only a last check. So you would spend maybe 10 minutes clicking around a bit there to see if really everything's really, really, really all right. But usually our development environments and our testing environments are close enough to production for us to be relatively sure that things will work the same.
Michaela: [00:35:51] And what about these testers that you said all of those tests are automatic tests that really end to end test, and those you would run before deploying to production as well?
Leif: [00:36:02] Yeah, they could run automatically for every pull request. So if you create a pull request that has a label "needs review", then that will automatically run all the end to end tests on it.
Michaela: [00:36:13] Cool, and what about technical debt? I mean, that's. A topic that's discussed hotly and it's a very fuzzy term as well, but I would like to talk with you about it, but role has technical debt in Automattic and how do you handle it? How do you assess it on a day to day?
Leif: [00:36:30] Right. I mean, we definitely do have technical debt. There's no way around that. Mostly because we try to move fast and we try to work only on the things that we really want to accomplish. And, that can be a aspect where, for example, future maintainability, can become more difficult because you're trying to optimize for shipping customer value. Right?
Once a quarter, we do have a week where basically every developer works on things like that. Either technical debt or also things like UX debt, like where where flows don't match with each other, but it's not a priority for any team. So it kind of lingers in there. But once a quarter, we have this week where we address things like that. And over, over the weeks we collect tasks that we think would be a good idea for such a week. And during the week, we will then have like a pile of tech debt or UX debt tasks that we can work on. Actually, we are in such a work. In, in such a week right now.
Michaela: [00:37:42] So then that's your main focus or your sole focus?
Leif: [00:37:45] Yeah, pretty much the sole focus. I mean, the lights need to be kept on, obviously. So you cannot 100% focus on that. Um, usually, but it's a large part of my day. If I'm not on a podcast, um, then I would usually try to work on those tasks. When the whole company tries to address tech debt and UX debt once a quarter for a whole week. That means you get a nice cadence into it. You get into a rhythm where things, things aren't perfect at any one time, but they continuously get better step-by-step you know? And I think that's how we try to address it. And then we can refactor things while we work on it. Right. So we try to build that kind of maintenance into our daily work also. But, um, it's not our sole focus, right? Like we, we try to optimize for shipping customer value.
Michaela: [00:38:48] Yeah. Yeah. I like the really scheduled and planned approach as well because otherwise people tend to work on things that are just more impactful or they you know, like to work on shipping feature. They somehow always feels probably more urgent or more fun as well. So it's sort of bookkeeping, right? So when tax season ends, you somehow have to do your taxes. Right. So if there wouldn't be a deadline, probably nobody would do that.
Leif: [00:39:15] Yeah. Or documentation. I recently saw a talk where she discussed the incentives around writing documentation, and she said that nobody ever got fired for not writing documentation, and nobody ever got promoted for writing documentation, which is probably roughly true. So what incentive is there to do these things, right? But if you see the whole company turn towards this one goal for a whole week, then again, You want to behave as most people do. So, okay. Let's look at tech debt this week.
Michaela: [00:39:48] Yeah. True. I'd also like that pick up the topic about promotion. So how, how do you think you're assessed? How is, is there a form of process for, is, do you know how people assess you? What they look for? How'd you get promoted? What's the life there that Automattic.
Leif: [00:40:04] Hmm. So what's the promotion, So when you like a change in title? or the change in the responsibility?
Michaela: [00:40:14] Yeah good question, maybe I don't know. Very often I think it's a change in title, better compensation for what you do.
Leif: [00:40:23] So, so a compensation is reviewed once a year automatically, you don't have to do anything. And basically it takes into account the work you did the year before, and depending on how impactful you were on how much value you shipped to customers, basically your salary will get adjusted, by more or by less. You'll never get a reduction. You'll always get a slight increase just to account for inflation, but that's all there is to compensation and Automattic.
Michaela: [00:40:54] But how would you assess the quality that you shipped or the value that you ship to customers? I mean, that's not that easy to measure I imagine.
Leif: [00:41:01] Sure. Um, so regularly your team lead will get a request for feedback and they will like try to document everything that you saw about your report and report that in that request for feedback and the employee themselves will also get such a feedback request for self feedback, and they can also give feedback on their lead. So, and it can be anonymous if you want to, you can decide to share that feedback with a person, but you don't have to.
And that's how we get a constant stream of feedback about people, but on the formal side and on the career track side, we don't have any titles. Really. We can choose our own titles. Let me quickly look that up. Matt Mullenweg with, I looked that up in the employee directory, his title is "C-BBQ-TT", which stands for Chief Barbeque Taste Tester, I think - something like that. So it can be very random, can be fun, but it can also be a matter of fact. So some people will say senior developer and that's what they are. So roll or responsibility, isn't tied closely to compensation. So if you become a team lead that we try to not consider that a promotion, there will be no change in compensation if you change your role.
The incentives are designed in a way to optimize for people who really want to do the things that the role requires. So if you become a lead and your salary doesn't change, it's suddenly much easier. Just say, wait, "Yeah, this is really annoying. Like I've given this a shot for a year" and actually, I like writing code better and there will be no change in compensation, so you can do that. So, and that way we try to keep the people in the lead roles, in the management roles who really like the management work right. At a traditional company, you would, I don't know, get like pay bump once you're a, like a manager and then you adjust to the new lifestyle and you can't go back and suddenly you're stuck with this role and that's something we try to guard against.
Michaela: [00:43:29] Okay. Yeah. I see that. And so you said at the beginning, you said that you're striving for a lot of transparency and things like that. Is that also reflected, for example, in the, I mean, especially if you don't have those titles, I think on one hand, it's really good because it takes away though on the other hand, it's a little bit less transparent, right?
Leif: [00:43:51] It's definitely a double edged sword. I mean, also just people coming in and it's not something that'd be advertised warning: We have no career titles here. Right. And that's something you definitely have to grow into. But to me it was, it is somewhat freeing, you know where to realize that you don't have to follow this career ladder that someone somewhere designed at some point, but that, and basically make your own adventure right. And grow in ways that you think are interesting for you at this moment in time, it is a huge adjustment, but if you're still here or if you're still at Automattic after two years, you will probably have adjusted to this way. And, if not, then you might think about leaving because I don't know, you crave external career progression more, titles more, maybe you're a very title oriented per person, then that would maybe not be a good fit because everyone should choose the random title.
Michaela: [00:44:50] Yeah, I can see how that fits into the Automattic culture and the idea of being humble. Well, one thing that I wanted to talk with you about, and that I think fits a little bit into that topic as well is, diversity. How diverse is, especially the technical workforce, at Automattic.
Leif: [00:45:09] Um, so the tech industry in general is basically white dudes, right? And we are mostly that too. It's something we are trying to actively address by being more present at, at events where a much more diverse audience might be present. We try to reach out proactively a bit more. We try to make ourselves available. We try to remove our own biases when hiring by for example, When we decided to hire a person, we have not have had a video chat with them. So you can go through the whole process before anyone sees your face. For example, we also do have an internal diversity and inclusion committee that meets regularly and tries to take care of where the whole company is headed regards to that, but it's not solved, right? Like it's very, obviously still a problem. And that's the case for us too. We're all very aware of it. And for example, the developer hiring is 50% or more than 50% females. And that that is really nice, you know, like, so the people who can make the decisions, the people who will impact what kind of people will join. They're already, at about 50% I believe. And I think that's a pretty good sign to have, like, they're kind of like the gatekeepers, right. Even though they try to be impartial and reduce their own biases. They, they are the gatekeepers. And it's nice to know that there is a, a larger number of perspectives on that team than there has been traditionally.
Michaela: [00:46:50] So, yeah, I can really see how that can impact the hiring decisions, and I think it's a really good step forward.
I wanted to thank you Leif, for this awesome interview. It was, I think, very well rounded. We covered a lot of topics. What do you think?
Leif: [00:47:04] My pleaseure. Yeah I think so, very nice.
Michaela: [00:47:05] So thank you for being on my show and letting me know so much about your work life at Automattic.
Leif: [00:47:11] Yes. Thanks for having me.
Michaela: [00:47:14] Thank you Leif talk soon. Bye bye.
Michaela: [00:47:16] I hope you enjoyed another episode of the Software Engineering Unlocked podcast. Don't forget to subscribe. And I'll talk to you again in two weeks. Bye!
Copyright 2022 Doctor McKayla