In this episode of the groCTO Podcast, host Kovid Batra interviews Anton Zaides, the Director of Engineering at Taranis and author of the Leading Developers newsletter. Their discussion focuses on the challenges and strategies involved in leading development teams versus platform teams.
He recounts how his early interest in gaming and experiences as a guild master in World of Warcraft shaped his leadership style, teaching him valuable lessons in social intelligence and teamwork. Maher outlines his proprietary framework for peak performance focusing on shared understanding, trust, and competence, and highlights the significant benefits of leveraging generative AI tools like GitHub Copilot for improving productivity. The episode also delves into the complexities of implementing new technologies and managing distributed teams, underscoring Maher’s strategies for overcoming these challenges through continuous learning and fostering a collaborative culture.
Kovid Batra: Hi everyone. This is Kovid, back with another episode of groCTO by Typo. And today with us, we have a very special guest who is coming to the show for the second time, but first time for this year. That’s Anton. Welcome to the show, Anton.
Anton Zaides: Thank you, Kovid. Great to be back.
Kovid Batra: So let me introduce Anton. Uh, so Anton, guys, is Director of Engineering at Taranis, a company from Tel Aviv. And, uh, he is also the author of Leading Developers, which is a trending newsletter, at least on my list. And he is having almost 18,000 subscribers there, writing some great articles we are really fond of at groCTO. So congratulations to that, Anton, and welcome to the show again.
Anton Zaides: Thank you so much.
Kovid Batra: All right. Uh, so today’s topic of discussion is one of the topics from Anton’s newsletter, which is ‘Leading Dev Teams Vs Platform Teams’. This was a very interesting topic. Uh, I read the whole newsletter, Anton, and I really found it very interesting and that’s the reason I pulled you off here. And, uh, before we like jump into this, I’m really curious to ask you a few questions about it. But before that, I just want to know, uh, how was your last year? How did 2024 go? What are your plans for 2025? So that we get to know a little more about you.
Anton Zaides: So 24 was very busy. I had my, uh, I had my first kid at the beginning of the year, so a year ago, and got promoted a month after that. So it was a year full of..
Kovid Batra: Super hectic.
Anton Zaides: Yeah! Hectic, career, family, and I think a small one would be in my, uh, first international conference, uh, back in September, which was a great experience for me, you know, like talking in English with an audience. So I would say lot of family, lot of career. And in the next year it’s more about family. I’m right now taking a 7–8 months break and I’m planning to work on my own thing. Early child education, mainly helping parents, children, like my own kid’s age. Just a bit of technology and also learn about it. You know, I feel parents don’t really know what they’re doing. So that’s my goal for next year, to be a better father and use technology for that.
Kovid Batra: No, that’s really amazing. I know this is, I think there are a few experiences in a human’s life and this is one of those which changes you completely. And, and in a, in a very good way, actually. Uh, when you’re young, you usually do not love to take responsibilities. Nobody loves to do that. But when such kind of responsibilities come in, uh, I think you, you grow as a person, there is something that, uh, something else that you explore in your life, at least I would, I’ve seen, uh, in my friend circle and of course, I can relate to what you’re saying also. So, congratulations and all the best. Uh, we really feel that you would do great here as well.
Anton Zaides: Thank you. Thank you. Definitely. We’ll try.
Kovid Batra: Yeah. All right, Anton, uh, coming to the main section, uh, talking about platform teams and dev teams, uh, this topic is very unique in, uh, in a way that nobody usually gets to talk about it in detail, in depth the way you have done it. Of course, a lot of generic articles are there. I’ve read a lot. This session could be a really good guide for someone who is, uh, in a position where they are moving into these roles from, uh, leading dev teams to platform teams. They could really have some learnings from what you have experienced in the past. So, first question to you actually, why did this topic come to you? What happened in your personal experience that made you realize that, okay, this could be something that an engineering manager or a tech lead who is switching between these kind of responsibilities would be interested in knowing?
Anton Zaides: Going back, I first started in a classic dev team, right? I wrote code like everyone else for a few years, and then I switched to the platform side, DevOps side, more infrastructure, and led the team there for a couple of years. And I decided to switch back. So it was two switches I did. And in my last role as an engineering manager of a classic product-facing, you know, user-facing team, I felt that most of the other engineering managers in the organization, they don’t really know how to work with the platform team. We have a DevOps platform team that provide us, you know, all the tools, they help us, and I felt they don’t really understand, uh, how to approach them, how to help them, how to connect them to the business. So they just really liked working with my team and I always got what I wanted and I pushed the agenda for that. And it really, really helped my developers too, right? Because they got close to the platform developers and they understood it better, that made them better developers. And I felt like this connection can help other engineering managers who never experienced how difficult it is to be in a platform or DevOps team. I’m using the terms, uh, interchangeably, but, uh, let’s call them platform for now. So I felt that, you know, I can show the other side and I hope it will help other engineering managers to see the difficulties and stop being annoying, because, you know, we are the, we are the clients. It’s very, very hard to satisfy developers for platform teams. It’s almost impossible. You’re always too slow. You’re always like, too many bugs. You’re always not prioritizing me enough. So I wanted to show a bit of the other side. So that was the focus of the article, like showing the inside of a DevOps team with some tips, product teams on how to help the, those DevOps teams. That was the idea.
Kovid Batra: Hmm. Interesting. Interesting. So this was some real pain coming out there and like you telling people, okay, this is what the picture is so that they know what’s going on. Right. I think that makes a lot of sense. And I think a lot of people connected to that. And even I like the article a lot. Um, I was reading one section, uh, from the article, which mentions about, like this is something which is really, really hard to manage, right? Uh, because the, the expectations are very hard and you just now mentioned about, uh, it’s, it’s very hard to satisfy the developers and then the requirements are changing too fast. So these were the first two things I remember from your article which were, you, you touched base upon. So can you just give me some examples and the audience about how you see things are changing really fast or how it is becoming very difficult for you to manage these demanding clients, actually?
Anton Zaides: First of all, I think when your clients are technical and they are inside the company, they feel the privilege to tell you how to do things and prioritize your work, right? Because they say, Oh, why does it take you a month? So, I know I can do it for a week, right? They feel they can do the platform work and they kind of push the platform teams. Um, I had an example where when I was doing the platform team, we were responsible for, I don’t want to get too technical, but we had, uh, you know, database services like Postgres, MongoDB, Redis, right? Storage databases. So we were in a private cloud and we were responsible for, uh, providing those database as a service. What do you have in AWS and GCP? You just can request one. So we needed to do the same in our own private cloud, which is quite complex. And we provided PostgreSQL and MongoDB and Redis. And every day another developer says like, why don’t you do Cassandra? Or why don’t you do CouchDB? Like they felt like they know what needs to be done and they didn’t. They never thought, you know, in my opinion, Postgres is perfect for 99.9% of the startups and their products, but the developers felt like they need to push me to provide them new database just because they wanted to use new technologies, right? And now I heard like, uh, for example, we have Jenkins, right? So in my company, I heard developers complain, why Jenkins? It’s so slow. We need to replace it for something faster. Right. And this is something as a product team, you’ll never hear your client tell you, why do you use React? You need to use Vue. Right? It’s faster. It’s, they don’t care, right? They care about the end result. And here the comments like this, like does somebody really know how hard it is to replace Jenkins with another tool? What are the costs? What are the benefits? Why do it? So So they feel very comfortable, like, suggesting and giving their opinion, even if nobody really asks them, I would say. That’s one thing.
And the other one about the priorities is it’s actually, I would say sense of urgency that there are a lot more fires in the platform teams. For example, if you have, uh, we had the case of a GPU problem, right? You know, the world has, uh, not enough GPUs. So we had, we use, uh, the cheaper version of GPUs where they don’t promise you enough. And then we had a bottleneck and we needed the GPUs, but we couldn’t get them. And now we needed to change all the infrastructure to request the higher GPUs and kind of balance them to save prices. And this is a project that took one month and it’s completely stopped what they’re working on, which was also important. And you have so many incoming things like that, you know, you have an alert somewhere, right? Something is crashing. Very often it’s the developer’s problem. But if you see, uh, prod crashing, you say, okay, it’s, it’s the DevOps. They don’t have enough memory or they don’t have enough nodes or something like that. And then you kind of need to debug and then you understand it’s the developer’s problem. You tell them and then they debug and come back to you because they don’t do their job well. So this all back and forth makes it very, very, very hard to concentrate. I remember sitting in, you know, you have this tap on the shoulder, “Please help me a bit. Uh, please explain to me why this is not working.” Uh, clients usually in a product team, you have customer support, you have customer success. You have so many layers that isolate the developers from distractions, right? And you can see it straight here. Your clients are sitting by your side and they just go over and sit by you expecting you to help them. I think product developers would have been crazy if your client would come up to you and say, “Oh, this. I see an error, help fix it now.” So, yeah, I agree. Those are the two things that, that make it, uh, very hard, clients being opinionated and always distracted.
Kovid Batra: Right. I think from the two points that you mentioned, uh, there is always unwanted suggestions, recommendations, and then there is, there is this explanation when you do not want to be directly interacting with them, there should be a first level of curation on whether the problem belongs to the platform team or to the developer, there should be some level of clarity there and then probably there should be deep diving into what’s going on, who’s responsible. So what I felt is, let’s say just hypothetically, uh, five years down the line, you are an engineering leader who is managing the complete tech for, for an org. Uh, you have platform team, you have your development team, right? What advice or what kind of culture you would like to set in? Because it seems like a problem of a culture or perception where people like blame the platform teams or do not empathize with the platform teams that much. So, as an engineering leader down the line who is leading two different teams, what kind of culture you would like to set in or what kind of practices you would want to set in so that platform teams who are equally critical and responsible and accountable for things as development teams are operating neck to neck? Or I’m not, I’m short of words here, but I hope you get the sense of what I’m trying to say.
Anton Zaides: Yeah, I think I got it and, it’s, it’s a small thing that we’ve actually tried, but I think if I would have been the decision maker to be on a biggest scale, actually to switch places for at least a while. So I believe that platform and DevOps knowledge is super useful for every engineer, right? Not always the other way around. So I truly believe that every product engineer should know about platform, at least the basics, not every platform engineer should know React, right? Depends on what they work in, but I would put the product engineers and put them for a month, uh, helping the platform teams in a project. Like, everyone should do a bit of platform work to understand, to see how they work, right? They can work in Kanban and not your usual scrum to see how they’re day to day. If you see from the other side, like if you need to provide support to your own team, right, you are the pipeline. You will see how many requests are coming through and the other way around. I feel that we had, uh, for two sprints, like for a month, we had one of the platform developers in our team because he wanted to experience the life of a developer to understand the problem better and the usage of his own systems. And it was really, really mind opening for him too, to understand why we complain, what he thought was so easy to understand that it’s our problem. Once he sat with us and tried and developed and, uh, released some backend code to production and understood it’s not that easy. And so this connection of switching places and it has some cost, but I feel it’s worth it.
And the second one I would say is connect, like the road map shouldn’t be different, right? They should be much more connected. So when you’re building the platform roadmap, you should have, of course, the engineering managers, but not only when you build it. Like, they should be there at every release kickoff, every, every time they should be part of the platform roadmap. This is the easy part. The harder part is to explain to the platform people the your product, right, how is your 3–4 months going to look? What are you working on? What do you expect? And not just the managers, which is what usually happens, right? You have a manager sitting with a manager, discussing and stuff like that. The people underneath need to understand that, uh, sit there. For example, a platform engineer should hear customer success stories that he indirectly helped because a big part of the problem that when you work in the platform team, you don’t really affect the business bottom line, right? You help developers create solutions, but if you can have those stories of how you helped someone deliver something faster and what was the impact on the company, it creates like a shared responsibility because next time you will want to help them faster. You will want to understand the problem better because you feel the impact. Saying, “I released the service to production in five minutes instead of three hours.” That’s nice. But saying, “I released the feature a week earlier and a bigger deal was, uh, agreed by the customer because of the DevOps team.” Right? Doing this connection. It’s not always easy, but in a couple of cases, we were able to do that connection. Um, platform work directly to business outcomes. I feel that would be something that we try, uh, much more. Um, so yeah, if I had to choose one, it’s just, uh, switching the places a bit, we had a concept called ‘DevOps Champions’, but it can be ‘Platform Champions’, uh, where you pick one developer from each product team and they have a weekly meeting with the platform team and like hear about the latest news, ask questions. And for example, they are the point of contact before you can contact the platform team. You have someone in your team who is interested in platform and he gets more, uh, he gets like, I would say Slack, direct Slack access to the DevOps team They know like this person, if you ask, we will drop everything and help them. And they, they do trust. And then the whole team talks to one person instead of to the DevOps team. And, and this helps a bit. So I hope it was not too confusing. So if I sum it up, I say switch places and have a dedicated platform, uh, representative inside the product teams and also connect the platform team to the business side. Yeah.
Kovid Batra: That really makes sense. Uh, this point which you mentioned about bringing DevOps Champions, right? Like who are going to be the point of contact for the product teams to share knowledge, understand things. Going back to your newsletter, uh, you mentioned about bringing more visibility and recognition also. So is this dev champs, DevOps Champions some way of recognition also that you want to bring in into the teams to have a better culture there? I mean, basically these teams lack that level of recognition just because they’re not, again, directly impacting the business. So they don’t really get to see or feel what exactly they have done is, is this an outcome or consequence of that?
Anton Zaides: No, I think it’s a bit different because the champions are product engineers, like who are originally from inside the team. So if I have five developers, one of them will be like, uh, will wear the platform hat, but he will be a product engineer and he will get to, to, uh, learn from them and work with them, the ones who are interested. For the recognition, I’m talking about recognition of the pure platform engineers, which are usually in the dark and separate there. And there it’s about what we, we discussed a bit earlier, also sharing their stories, but also public acknowledgement. That’s something that I really, I have the privilege of having a LinkedIn, you know, and I constantly write there. So I, I did a couple of shoutouts for our platform engineers after nice projects, and they really, really appreciated it because, you know, people usually, you know how it is. If it works, they don’t hear about platform, only when it breaks. So they don’t get like kudos for nice projects and stuff like that. So I really try both on LinkedIn, but also in internal companies like channels, you know, saying nice words, uh, appreciating the work, stuff like that.
Kovid Batra: Makes sense. Makes sense. Totally. I think, uh, one thing I would be interested in knowing, like any of the projects that you took up as a platform team lead and completed that project. What was the mindset, what was the need, uh, and then how you accomplished it? Just deep diving into that process of being a platform team lead, uh, leading a project to make the lives of your developers, uh, better and maybe making them more productive, maybe delivering faster.
Anton Zaides: So let me think, it’s been a while, right? It’s four or five years ago since I was there. But I think if I go back, right, my team’s role was to deliver database as a service for our customers, right? Customers and developers, they want, uh, whatever PostgreSQL, uh, MongoDB and they, it’s hard for me to explain to people how it is without a public cloud. I was in a government agency, so there was no GCP, AWS, Azure. It was like everything, you need to create everything. It was an air gapped environment. Because of, you know..
Kovid Batra: Uh, information, regulation.
Anton Zaides: Regulation, information, you couldn’t use stuff like that. So we need to do everything from, from scratch. And one thing that, uh, we were a small team, so all the communication was, uh, we didn’t have like a portal, right? I know it’s very hard to imagine a world without the public cloud, but it was like emails and messages, please create me a database and stuff like that. And one very small annoying thing was the extensions and Postgres. You have many default extensions, like you have PostGIS, like for geographic extensions, you have like, uh, for using it as a vector database, you have many extensions, and we wanted to help them use those extensions, right? Because every time they needed a new extension, they need to send us an email. We need to check it. We needed to roll it out and stuff like that. So I know it’s, I think it’s not what ideally what you, uh, meant because it was quite a small project, but I saw that pain and we kind of went and figured out the top 20–30 extensions that did some templates and did some UI work, which is quite rare for platform teams, right? Because you hate UI, usually if you’re in platform. At most, you can do some backend, but you prefer to do like, you know, flash scripts and stuff. So we did some basic, uh, interface with React, HTML, CSS. So to create this very ugly portal, which I think people appreciated. It makes the work easier. And I think the good, the good platform teams are not afraid of writing a bit of code and using like graphical interface to a small portal or like, uh, if you want to request to see stuff like that instead of waiting for product teams to help them create a nice screen and stuff like that. Now with Cursor and, you know, and all the LLM, it can take you 30 minutes to do everything you need. Like, you have APIs, you can put them where they can use buttons to do like that, you need to request something. So I think like that barrier, if I go back to the story to break the barrier and not say, okay, I can only do backend stuff. That’s how it works. I will. And just think about the next step and go where it’s, it’s uncomfortable. I had, I was lucky because I had the background as a product developer, so it’s easy for me. But all of my team members, there was like, no, no way we’re going to write React. No, it’s not our job and stuff like that. So I had to, to force them a bit, force them and I actually enjoyed it because you know, it’s It’s, it’s rarely in the platform that you can actually see something immediately
Kovid Batra: This was an interesting experience and how this experience would have changed in case of such kind of requirement when it comes to dev teams, like, because we are just comparing like a while leading dev teams is different from leading platform teams. So in this situation, of course, there was a barrier. Uh, there was a problem which the platform teams had to solve, but it came with a solution that platform teams are usually not inclined towards like building the UI, right? If a similar kind of a situation had to come in for the dev teams, how do you think it would have been easier or difficult for you to manage as a manager?
Anton Zaides: I would say as a dev team, you have a product manager, you have UX designers, and you get a ready Figma of how it should look like, and you just implement it in, in a couple of days, right? It’s so much easier because someone is doing the research of talking to the customers. Some platform teams have a product manager, right? I would not say, but they for sure don’t have a UX designer working with them, because the system is internals and everybody say, “Oh, just make it good enough. Uh, these are our people anyway. You don’t need to make it beautiful.” So this, this is usually how it works. And in the product team, for me as a manager, it’s so much, much less work for me. The product manager, uh, doing most of the work. And I would just like, you know, manage the people a bit, coach them. But as a platform team, I did it, like 50% of my job I did product management. For some of the time I did have a dedicated product manager, but some of that I didn’t and I needed to kind of fill the hole myself. Yeah, because in platform team, it’s the first team where you cut the product manager. You say, “Oh, it’s internal. No need. Uh, the engineering manager can manage.”
Kovid Batra: That’s even my point, yeah. So even I, I felt so, like for platform teams, do you think it is even important to have a product manager? Because the tech lead or probably the engineering manager who’s involved with the team would be a good start to make sure like things are falling in the right places and understanding the problem. See, ultimately for a product manager, it is very important to be more empathetic towards the client’s problems and be able to relate to it. The more they relate, The more fit is there, the better solutioning they can design. Right. Similarly for an engineering manager who is leading the platform team, it would be more of a product role and it makes more sense also, as per my understanding. What do you have to say about that?
Anton Zaides: I have had experience with product managers with platform team who didn’t come from an engineering background and it was always a failure in my experience. Uh, I would say it’s better to have no product manager to let the engineering manager do the job. And ideally in, in that team after, I think it was after a year and a half, one of the engineers, like she mentioned she wants to become a product manager. This is her career path and then it’s a perfect fit, right? If you have an engineer who wants to become a product manager from inside the company, then it can work great. But I feel that in the platform case, the product manager must have an engineering background. Otherwise, like you can try to learn to be technical, but it would just be, it would be a different language. It would be, it’s not like product teams. Yeah, I agree. I feel it’s, uh, yeah, it just doesn’t work in my experience.
Kovid Batra: Makes sense. By leading a platform team where you find this kind of a fit where some engineer who is interested in becoming a product manager comes in and plays a role, I think I sense that there is definitely a need of a person who understands the pain, whether that person is an engineer or the engineering manager who is working as a product manager, but you definitely need that kind of a support in the system to make sure that requirements are flowing in correctly, right?
Anton Zaides: Yeah, I agree.
Kovid Batra: And most of the time what I have seen or felt is that engineers usually shy away or the engineering team shies away from being involved that aggressively towards client requirements. So when it comes to platform teams, how do you bring that extra level of empathy towards customer problems? Of course, they are developers, they relate to the problem, but still, I feel that in a world where we live dealing with real world problems, being a developer, you still get to see some side of it because you’re a human, you’re living in the, in that world. But when it comes to platform teams, it’s all technical. You have seen things, but still, it’s more like you are just solving a technical problem. So the empathy towards deep diving into the problem and bringing up a solution, does it become harder or easier when you are raising a product manager in an engineering team for platform teams?
Anton Zaides: I think it’s quite hard and I think this is the role of the engineering manager, of the platform engineering manager. Like I feel the product managers still have difficulty bridging that gap. I would say that platform engineers, either by experience or by character, they care more about the technical side. You know this term of product engineer, which is like pure product engineer, not like software engineer, like the people who decide what to build. Platform engineers, from my experience, care about the technical side, like much, much more, right? They want to build excellent solutions, they are excited by crazy bugs and they are excited by saving costs, stuff that most people are less excited by that. And yeah, it’s, it’s purely the job of the engineering manager. Like, as a platform manager, you need to show the pains of the developers too. That’s much more than in a product team where the PM filled that gap. I feel that even if a PM is an ex-engineer, in my experience, somehow, like, if the engineering manager won’t do it, the developers will resist much more the PM. Right? I think that’s what comes to mind. You have much more resistance in the platform team because they want to stay in the code. They don’t want to join customer meetings. They don’t want those things. Just want to code. So you need to, you know, like, uh, peel the shell and try to bring developers to share their stories, send them for a month for a development team, as we discussed, which they will hate probably. So you need to, to, push a bit. And the PM, it’s not, they are not his or her direct report. So they have limited power and you can actually, I would not say force, but kind of help them hardly along that path, uh, of understanding the user brains. Yeah.
Kovid Batra: Great, Anton. I think, um, thanks. Thanks for this interesting talk and helping us deep dive into the platform teams and the dev teams and how they differ in their core DNA. Uh, I think there were some great insights about how things change when you are leading a platform team, that from the expectations, from the kind of mindset that the developers come with, the unwanted suggestions, and like how you bring more connectedness to the business and recognizing teams. So I think this was a very interesting talk. Before we moved from the session, uh, is there any advice, uh, parting advice that you would like to give to the audience?
Anton Zaides: My main advice would be to the product leaders, product engineering managers to try much harder to understand the pain of the platform teams in your organization and how can you help them. Schedule 1-on-1s with the platform engineering manager, be more involved because they will appreciate that help and they might not even know they need your help. And in my experience, you will benefit for sure.
Kovid Batra: Makes sense. Makes sense. I think this would not only help reducing the friction, but will also help, uh, in bringing a better and a collaborative effort to build better product also like better platforms also.
Anton Zaides: For sure.
Kovid Batra: Great, Anton. Thank you. Thank you so much once again, uh, it was great having you on the show. Thank you.
Anton Zaides: Thank you, Kovid. It was great being here.
Implement engineering metrics, gather SDLC insights & set continuous improvement goals in your dev teams.
Try Live Demo
