This is a recap for myself so that I can remember 2-3 years from now what I did in 2025.

  1. I began the year by achieving my first AWS certificate, the AWS Solution Architect Associate. It took quite some time to prepare and the exam itself ended up being a fun story to tell for my close friends. I scored pretty well on the exam even though I was pressed by time (not only by the usual time-constraint). So the year got off a great start.

  2. Soon, I was asked to join a project as a Python Backend Engineer for an undisclosed american client. The project itself was really interesting: domain was great, technical challenges and the tech stack were also great. In short, we built an AI-integrated web app which was mostly parsing user uploaded documents.

    1. This was my first, pure fullstack project when I was working on the backend team. It was quite an interesting experience, one that maybe should have come earlier, considering I’m on the ‘Python-track’. Nonetheless, it was great to learn from experienced full-stack and backend developers while also spending more time around american clients and their pace plus way of working.

    2. I very vividly remember I was responsible for our sync to async migration on SQLAlchemy and I got to debug a bunch of greenlet errors.

    3. I was also involved in creating a sharing system between users and teams for the app, something which I did not see the end of as the project had to be downsized and I was the easy casualty as the late-joiner (plus I was staying on for much longer than initially agreed).

    4. This was also my first time when I got to play around with LLM tools in a production environment. We were given access to Cursor which was a very useful experience even though back then MCPs were not yet a thing.

    5. Tech stack was FastAPI, Databricks (including Apps), SQLAlchemy

  3. After the project finished I spent some time on looking at possible tools we might use at future projects: I looked into an AWS data tool that was fairly fresh, but looked unproven (funny thing, it ended up being my verdict as well) and I helped in the development of an inhouse AI tool. The former felt more like a grind cause I had to debug and try to somehow make it work, while bashing my face against different errors whenever I progressed from the previous one. The inhouse tool was fun, I got to write up a mini tool that helps the migration of Snowflake tasks to Databricks, I wrote up an article on that which can be found here.

  4. To wrap up the year, I joined another ongoing project where we had to help the client with hail damage predictions. Very interesting domain, simple tech stack containing MSSQL, SQL and Python. This was my biggest learning of the year in terms of how to manage projects and what I’d do differently in hindsight.

    1. As I got onboarded I noticed a couple of issues, things that I would’ve done differently but are still fine as-is. I did not raise them to the team, as we were already running late on the project and figured there would be no chance of us getting time on work on them. In hindsight, I should’ve just collected them, send them to team and let them decide.

    2. As the project was running late and part of the team was going away on holiday some of the client related communication rested on my shoulders. It was a nice change, as I usually don’t have to do this.

    3. A very notable part of this project experience was when we had to do an optimization on one part of the pipeline. Due to time restriction, our CTO got involved and handled the whole problem like I’ve never seen. It made me feel like I’m in the wrong business, but let’s just say (or hope) he’s the CTO for a reason and I only need some mileage to get there.

    4. Things I learned: geopandas, numba, python multiprocessing, optimization (and asking good questions to get the best output)

  5. During the year I also got a couple of extra responsibilities within our data team (mostly due to an in-house merger).

    1. I started to co-lead our inhouse knowledge sharing sessions. This was already on my plate unofficially, and it made sense to join forces with a colleague who’s been running it for one part of the team as well.

      1. Sadly so far the attendee numbers are not as high as I would’ve hoped, even though we’re bringing interesting topics and presenters. Another pet peeve of mine is that the amount of feedback is quite lackluster from the team, so I have a hard time understanding why people seem to be not so interested.

      2. I’m involved in helping the teams and developers on how to use AI, raise awareness and help those who’re unsure what’s the best way to do it. There wasn’t too much happening on this front, but it feels nice to be involved as it’s close to my heart.

  6. Slightly connected to the inhouse knowledge sharing sessions, I've also done 2 presentations to the team this year.

    1. First was a technical one on AWS Sagemaker Lakehouse while the second was still aimed for developers, but in a more general and less detail-heavy tone: I recapped what people might have missed from the last half year from the tech world if they didn’t keep up and paid attention.

    2. I (would like to) think that usually I present much better, but for some reason I don’t think I was able to bring the same form and energy into the presentations like in the previous years. Maybe the fact that the audience was different or that the I didn’t have too much time to prepare for them paid its price, who knows.

  7. To wrap up the year I was involved in a small tender rfi response and task estimation, which was something I’ve never really done before, at least not to this extent. Not sure if it’ll be my favourite thing to do down the line, but I definitely need a bigger sample size to say that with confidence.

Overall it was an interesting year with ups and down, lots of stretching, lots of learning. And that’s always good.