I pretty much skipped doing a retrospective on how year three went in MSFT. If you read my previous ones (first and second), you’d have remembered that I spent year one waiting to be let go, and then I spent year two trying to understand why I wasn’t let go all the while trying to keep my head above the water for long enough to at least make it two years because I figured that two years in Big Tech is a good number to have on your CV. Then in year three I figured that probably I wasn’t that bad that I was on the immediate let-go list, so I worked even harder to ensure that, you know, I stay for at least three years because it’d just suck not to have three years of same company on my resume. Then something happened – at a full three year mark , I actually started to feel comfortable with my work and I felt like I finally understood all the pieces that make this giant machinery called Azure SQL. So let’s just say that at full three year cycle, I started feeling relatively comfortable. And now, as I’m a month shy of making a four year’s journey, I have some thoughts to share. Bear with me.
One thing that I genuinely knew nothing about is that, in Microsoft, it’s completely normal that you never have to leave the team for the charter that you are working on to change completely. And to make this even more amusing, in the very early days (as in – literally the first week) I happened to stumble upon the then-director of our Dev Center, Dragan, to whom I properly introduced myself with “Hello” and “nice to meet you”. You know, the usual sutff. It took me almost four years to fully grasp his completely cryptic response: “Don’t get used to anything. Everything keeps changing.”. No “nice to meet you”, no “welcome aboard”, no “cool!”. Just pure, simple and fully mysterious response. He was, in fact, right, which I was yet to learn.
I spent 3+ years being part of a team that dealt with Identity-related stuff, Permissions-related stuff, Transparent data encryption related stuff, Windows Authentication related stuff, … essentially, everything that could be considered as Security-related is what we dealt with. Amusingly, people occasionally referred to us as an “AKV Team” because majority of stuff that we dealt with ended up being related to Azure Key Vault (AKV). Because, if you are using Azure SQL DB or Azure SQL Managed Instance, and if you want to bring your own encryption keys to Cloud, Azure Key Vault is the way to go. But there are so many ways these setups can go avry that, as I said, we occasionally were called “AKV team” due to sheer amount of AKV-related stuff we had to deal with.
This whole “AKV Team” joke escalated so much so that at one point, when Claude Sonnet 4.5 arrived, I actually spent almost 3 full monts of my FREE TIME (nights, weekends, etc.) building a so-called “AKV Mock”. This was essentially myself vibe-coding a clone (or, a Mock/a Stub if you will) of real Azure Key Vault, in order to understand how hard would it be to make one that can be used with real SQL Server. Hint: three months of on&off free-time work and vibe-coding 🙂
Slide from the internal talk I gave on AKV Mock
You might be wondering why the hell am I talking about some random shit – the “AKV Mock”. Who cares, right? Well, great that you asked – mostly no one cared; except for myself! As I said before, the three year mark in MSFT was when I started feeling COMFORTABLE and that also happens to be around the time that Sonnet 4.5 came out which is a combination of facts that fused neatly together to reignite that old and forgotten spark of building cool shit. And as a member of the team that had to deal with AKV-related incidents, what else might one want more than building a clone of the exact same thing I worked on/around for most of the day! It reminds me of this meme:
My point is that at year three, gaining an understanding of what the whole Azure SQL machinery is all about, coupled with the coming of AI models that could actually code stuff that doesn’t shit the bed on its first execution, I went into a full-on frenzy mode of BUILDING. Not necessaly building stuff that anyone asked me to, but rather the stuff I fucking loved building. And let me tell you, I absolutely LOVED every moment of building that AKV Mock thing! I was going to bed late and waking up early, stealing every possible free moment that I found to “add just one more thing to it” or “solve just that one problem that keeps popping up”. It was insane and energetic and I loved it.
Now mind you, the missing piece of this puzzle, and the reason why I’m writing this whole article is that, at the time, I also have another pet project living rent-free with me – an adorable and loving three year old little devil who required daddy’s attention more than ever. But we’ll get to that.
Right about the time I was wrapping up a v1 of the AKV Mock, and as I was prepping to go into the v2 adventure of it (again, no one asked me about it and no one forced me to do it; but some people found it really cool (thanks Rajat for loving it and offering even to fund further work on it!)) I received the new news – starting with year 2026, our (Serbian) team is changing charter – we will work on the SQL Extensibility area. Remember that cryptic “Everything keeps changing” response from before? Yeah, turns out that he was right.
For those not knowing what Extensibility is (and trust me, I’ve heard all versions from “accessibility” to “building some boring SQL extensions”) – it’s one of the coolest things that a person who loves Databases can get to work on. Extensibility is a Platform that allows you to extend SQL Server’s engine with custom languages and libraries. We currently have open-sourced extensions for Java, Python, R and C# which means that you can build your library in any of those languages and use it alongside SQL by utilizing the Extensibility Architecture. Yeah, it’s like the coolest freaking thing you can get to work on, seriously! So, around February 2026, I moved to Extensibility Area.
This whole move was interesting because up until then – I did mildly-amusing things for regular salary and then I built AKV Mock in the off-hours for my own amusement. But with a move to Extensibility, I actually got a chance to be paid to build cool stuff in the working hours. This, trust me when I say it, is both a blessing and a massive curse, as I will elaborate next.
This is where I just can’t but not mention AI because it plays such a vital role, for good and for bad. In the, what I refer to “past era of software engineering”, onboarding to completely new and massively unknown area, would likely take at least 6 months, probably more like a year. Hell, if you recall what I said in th beginning, it took me ~2 years to start feeling comfortable in the Security area. But the LLM toys changed all of this, making navigation through massive landscape of unknown codebase just a prompt away. Instead of spending days&nights setting breakpoints to see how stuff works under the hood, now you could literally just ask. And coincidentally, we also got our internal MCP that “knows” our massive codebase, so it quite literally WAS all a prompt away. “Where to put a breakpoint to capture a Launchpad connection arriving to SQL?”, “How does Data Chunk message look like?”, “How and where is AppContainer constructed?”. Those are all questions that’d take months to digest and investigate in the old days, but now they were, quite literally, a prompt away. And this is important. For good and for bad, it’s freaking important. It’s a blessing and a curse, as I said.
Suffice to say, I became obsessed with Extensibility. I still loved my AKV Mock but I started treating it as a past relationship – we had fun and some good sex, but it was never about to become a thing. It was cool while it lasted but didn’t have that big potential of becoming something serious. Extensibility was a wife-material. Mostly because I found it fun to be around and I was, you know, actually paid to build that stuff. I literally joined and T-0 moment and was immediately tasked with building some really freaking cool new core features for the vNext platform that was being birthed. And this consumed me.
At first I was thinking about this during the day. It was just too much to digest, even with AI, so I was mostly thinking about it during the working hours. But then, as it started crystalizing and as I started seeing first results shortly after that, I started thinking about it after work as well. And then I started thinking about it BEFORE the work as well. So much to build. So many features to deliver. So many interesting engineering problems to tackle! So much that at one point I found myself waking up in the middle of the night, not even knowing if I was asleep or not, just thinking about the stuff that I was working on and stuff that I want to build on top of the stuff that I was already doing. And all the potential that this new thing that we are building carries. All the cool stuff that OTHERS can build on top of it. All the ways that people could extend SQL Server’s functionality with the stuff that we are engineering!
Worst of it? No one was pushing or forcing me. Not even once! Sure there were some deadlines and commitments, but did anyone force me? NO! And I think that’s the problem – if I was FORCED, I’d probably not enjoy as much and in retrospect, likely wouldn’t become obsessed. And yet, here’s this mega interesting thing, likely the most interesting thing I ever got to work on in my career (and mind you, I was actually being PAID to do it), and you’re telling me I’m supposed to keep my work-life balance in check? Hah!
But here’s the kicker – remember that three year old little dude living rent-free in our home? Yeah, he was now turning four. And by all accounts, he probably needs his dad now more than ever. And I was aware of that but I just couldn’t get my mind off all the things I need to build in Extensibility, getting to the point that I spent half of my vacation being awake during the nights thinking about the problems I was trying to solve. And now that I put it into words – yeah, I guess that sucks 🙂
Now I have to start a side-plot of this whole story which, as you will see, complements and merges into the main storyline in the end. As usual – bear with me.
As much as I would refuse to admit, aside from your parents, friends and later – your chosen partner, people who really may make or break both your personality and a career are your managers. Bad ones can really fuck you up, both career and mental-wise, good ones can really shape you into an amazing product, whereas the neutral ones can really steer you in any direction. But yeah, these are the people who really can influence you on so many levels. I was fortunate enough that the ones that managed me (and trust me – there was a LOT of them) were either on the neutral or on the positive side. This was really lucky because if I had an asshole-manager, my career and life likely would have had taken a different trajectory. But it didn’t and I consider that I kept growing year after year, which I can directly attribute to all the people who managed me. But some were actually exceptional. One really shaped me in the early days and my early twenties, when I was a hot-shot youngster, and that person is now one of the co-founders of a first Serbian unicorn, so you can imagine that we had quite matching energies. Another one retired in his early 40s, is living purely off of his investments and pretty much is living the life with his wife and three kids. This guy influenced my early and mid-30s, taught me the exact things that I missed and absolutely heavily influenced the trajectory of my further life and career. Oh, and I named my kid after him, so go figure. This is all not to say that others weren’t absolutely fantastic but it just so happened that two folks happened to be at the right time in the right place to guide and manage me and to shape majority of who I am today. And this finally leads me to the PRESENT DAY.
I was having a regular bi-weekly 1:1 with my current manager and we were discussing some of the latest happenings in the area that I’m working on. I worked extremely hard in the last weeks to try to meet some dates and interestingly, I actually failed to deliver, which sucked big time. And the convo was more or less about me explaining why I think that what we are working on is really freaking important and why I think it’s reasonable to work long hours, while my manager was trying to explain that yeah, but it’s still work and you need to balance it with life, especially when you have a kid. And then she proceeded to share a story with me. A story of one of her past managers who also worked really really hard because the stuff he was building was super exciting and important for the greater good (just like mine is :P).
I’m paraphrasing the story but the way this guy, let’s call him Bob, told it to my manager is – “So I worked really really hard to meet this project. Because it was important project and I enjoyed it. We had daily meetings at 7PM every day to match the Redmond working hours. And we kept it month after month. It was hard and I was sacrificing some family life for it, but it was important job that needed doing. And then, one day, my wife comes with my kid to visit me at the office, and my kid asked “Mommy, is this where daddy lives?””. And my manager proceeded to finish it with “and this is when I, being a highly-motivated youngster, realized that, as much I admired Bob and what he has done, I never want to be the Bob to my kid”.
And you know what? That resonated with me HARD. Like REALLY HARD. Because, coincidentally, just some weeks ago before the vacation, my son asked “Mom, is daddy going to work on the vacation too?” and my wife literally had no idea what to answer to that. And you know what? Daddy DID work on vacation and not because he HAD TO, but because daddy WANTED TO, because daddy considered this to be the really important project that he must deliver, completely forgetting about another long-lasting project at home – the kiddo who keeps seeing his daddy behind a laptop.
So I made a decision I’ll try to stay commited to – I’ll make sure to work from office most of the days and bring out the laptop only if stuff is really urgent (or during the on-calls, which is unavoidable). Because I don’t want to end up being the Bob, guy whose kid is going to ask whether daddy lives at his work. Because daddy DID start living with his work but I don’t want to be the one having a conversation with my future-son when he is in his 30s, explaining to him that daddy worked a lot because daddy loved his job and found it ultra important and exciting and even though I missed quality time with him it was all important because …. well, because I have no freaking idea WHY. It’s not. Especially not if no one is ASKING me to work more, which happens to be one of the major points that my manager was trying to bring. Which means daddy WILLINGLY CHOSE to work more over spending time with his son. And that makes daddy a shitty daddy, if you ask me 🙂
If you relate to this article and if it kicks anyhting in you, just try to remember that you, just like I, don’t want to be the one whose kid is asking their mommy: “Does daddy live here?”.
P.S. Whose cutting thiose f-kin’ onions?!