2025: The year of change and context

Where should I start? A LOT has happened within the last 12 months in 2025. Not only my job position changed and what I do, but also HOW I do things. But… let's start right from the beginning.

How it started

At the beginning of the year, I was employed as a DevOps engineer, and I really enjoyed the projects I’ve been working on. Sometimes I was writing code, sometimes “hunting down the ghost in the machine” (that’s how we called troubleshooting mysterious bugs), and sometimes planning and creating architectures for new projects.

Well..., 6 months later into the year, I handed in my resignation. Why? It’s simple: because of AI. Obviously, it’s AI - the thing all of us deal with these days. Some more than others, but everyone is interacting with AI in some way. Of course, there’s more to it than just “AI”, but that was the initial catalyst for the change.

I’ve been playing around with ChatGPT and tools like Cursor extensively, and the more I used those tools, the more I realized the potential of it. I was (and still am) amazed by what’s possible with them. Even with arguably “bad” LLMs (for today’s standards), used by simply copy-pasting code in/out of ChatGPT, I’ve been able to build many useful things faster and easier than ever before. Realizing how much is possible with those tools caused a mindset shift for me.

I think the way Geoffrey Huntley puts it here describes it perfectly:

The future belongs to people who can just do things

So yeah, I decided it's time for a change. I wanted to work with a small group of talented people where things move fast. Build, ship, break, fix. (Hopefully in a better order :D)

The change

Besides AI, one of the things I'm really into is observability, in particular OpenTelemetry.

OpenTelemetry brought me in contact with Juraci. We knew each other from the CNCF Slack, and at some point we met in person at an observability meetup in Berlin. We had great conversations about OTel, software development, and also OllyGarden.

OllyGarden had been incorporated for just 1-2 months at that time, but the conversation with Juraci instantly got me hooked. Not only because I immediately saw the potential and possible impact OllyGarden can have, but also because of the way Juraci’s opinions and mine aligned about topics like software development and AI.

We stayed in contact, and I told Juraci about my interest in working closer with him. Fast-forward from that moment on: I’m an official part of OllyGarden now.

Switching from my previous company, Clario, to OllyGarden wasn't an easy change. I met many great people at Clario (which I’m still in contact with as of today), and we’ve been dealing with exciting challenges within the DevOps team. However, due to the mentioned mindset shift, I decided to gravitate more towards the software-engineering direction in order to be able to “just build things”.

Looking back, I can say it was the right decision for myself. We have an amazing team here at OllyGarden, and in a world where more and more code, data, and telemetry is generated every day, our mission to reduce noisy and low-signal telemetry data becomes more important than ever before. I won't go into detail about it in this yearly review post, but I might cover the topic in a separate post in the future.

Honorable mention

I couldn't do a yearly review without mentioning Claude Code. I already touched on some of the tools I used in 2025, such as ChatGPT or Cursor, but honestly, the impact Claude Code had is on a different level. It disrupted the way I wrote code, fixed bugs, or hunted down the ghost in the machine (=troubleshooting). The majority of the code I’m shipping these days is completely generated by AI.

While (still) many people might think this is something bad, I don’t think so. Instead of code, I write specs. I define the contract, and the LLM implements it. Every change gets reviewed carefully by me. Of course, sometimes I might need to do manual adjustments, or sometimes, when writing business/performance-critical code, it still requires manual effort. However, I’d argue that most (not all) of the usual software development problems can be solved very well and much, much faster by LLMs than humans.

Highlighting such a tool this explicitly feels a bit exaggerated for a yearly review, but considering how much it influenced the way I work, I believe it deserves that mention. In fact, I see many smart people out there reporting similar experiences in their yearly reviews - apparently I'm not alone.

2026?

2026 is gonna be exciting. At OllyGarden we have many new features in the pipeline, which will be released in the upcoming months (stay tuned for KubeCon Europe!). Besides of that, I’ll constantly strive for additional optimizations in my software-development workflow. I find a lot of joy in experimenting with new tools or building them on my own. In terms of AI nobody really knows yet what 2026 will hold for us. At the beginning of 2025, AI agents weren’t really a thing, and now almost everyone is using them. New terms emerged: Context-engineering, vibe-coding, or You’re absolutely right ;).

I couldn’t be more excited to find out what will change, emerge, and be explored within the next 12 months, and what I’ll write about in the 2026 yearly review, but one thing is assured: It’s gonna be fun!