About me

I'm Peer Schlieker
an Applied AI student at Offenburg University. This fall I'll be joining GEOMAR for an internship semester on the MOLA project (Modular Ocean Lander). Before that I spent nearly three years as a machine learning working student at happyhotel, working across the full ML/AI/Data engineering stack: data pipelines, MLOps infrastructure, prompt engineering, production deployment. I'm focused on solving problems connected to protecting the natural world.

How I Got Here

I spend time outdoors. Sailing, bikepacking, hiking through forests. When you're regularly in nature, you notice what problems matter. For me, that means AI applied to ecological challenges.
That focus shaped my decision to build this portfolio and pursue skills deliberately. I'm learning Rust and GIS because they'll matter for the work I want to do. I also explore things like n8n and Optuna when they catch my attention, sometimes curiosity, sometimes because I see the potential. Coursework covers fundamentals; these are the skills I chase on my own terms.

What I Actually Do

At happyhotel, I worked across the full ML/AI/Data engineering stack for over two years. My first major project was a demand forecasting model using market indices. It went into production and ran for a while before we built a new meta customer model as a team, folding in a lot of what I'd learned from the original. Building it taught me how much weight sits in the decisions before training even starts: the architecture you pick, the features you choose to trust. After that I worked on infrastructure: migrated database schemas from dictionaries to Pydantic for type safety, built prompt engineering for a graph creation assistant, refactored metrics pipelines, worked on MLOps. The stuff that doesn't get noticed but keeps everything running. Next I'm at GEOMAR, on the MOLA project. What exactly I'll be working on is still open, could be audio embedding classification, could be earthquake monitoring. Either way it's ocean lander data, and I'll find out soon enough.

How I Think

I'm drawn to collaborative problem-solving. I work best when I'm thinking through a problem with others: whiteboarding, questioning assumptions, iterating toward a better solution. Code reviews uncover things you completely overlook on your own and push the quality of the work to a different level.

I also understand that meaningful work requires patience and iteration. Code you were proud of half a year ago often needs another look once you understand the problem better and your own skills have moved on. Sometimes halfway through a project you realize your approach needs rethinking. I don’t see that as failure. Adjusting your approach is far better than ending up with something no one can actually use.

What's Next

I'm building a career at the intersection of technical depth and ecological purpose. Right now, that means forest health analysis using satellite data and clustering algorithms. My path will probably take me deeper into applied environmental AI, research, or something unexpected.

If you want to see what I'm working on, check out my projects or GitHub. If you have questions, feel free to reach out.