LDPC Decoding — Master's Research at Nokia
Master thesis at TUM's LKN in partnership with Nokia Strategy & Technology. Simulated how environmental factors affect LDPC decoding.
I'm the engineer who keeps showing up in new domains.
Autonomous robots at 15. Four years at Infineon on automotive microcontrollers. Master's research at Nokia. A RAG hackathon win at EPFL in between. I'm looking for my next role, ideally one where I get to keep learning out loud, with AI and LLMs somewhere in the mix.

EPFL, one weekend, team of four. It was the peak of the LLM hype and every other team was gluing a chat bubble onto the Swisscom brand. I pushed ours the other way: instead of a bot, build what looks like a real new page of the Swisscom site, same layout, same type, same voice, with a single question field where an FAQ would sit, answering in article form as if it had shipped from their own CMS.
Under the hood we indexed the full Swisscom site into Pinecone, retrieved the top relevant chunks per question, and had GPT synthesize a grounded, article-length answer with links back to the exact source pages. Every test question the Swisscom team threw at it came back correct and properly sourced, and the fact that it felt native to their brand, not bolted on, is what caught the judges off guard.
I was on two hackathon teams that weekend. This was the one that won.
A few more projects across research, classical CV, and embedded hardware: the supporting evidence for the pattern.
Master thesis at TUM's LKN in partnership with Nokia Strategy & Technology. Simulated how environmental factors affect LDPC decoding.
Bachelor thesis at TUM's Microrobotic Bioengineering Lab. An automated counting system for microencapsulated cells built on computer vision. End-to-end from raw microscope frames to reproducible counts.
Built four autonomous competition robots with a fellow highschooler. Reached the semifinals in 2016 as the youngest squad in the competition. My first serious taste of embedded C under a deadline.
I grew up in Sousse, Tunisia, building autonomous robots with a school team. I went to TUM for electrical engineering because I wanted to understand systems from the transistor up, and I stayed for a master's. In between I spent almost four years at Infineon on automotive microcontrollers and wrote my master's research at Nokia on LDPC decoding.
I'm not loyal to a single stack. Over the years I've shipped classical computer vision, embedded C, React and Next.js, a Unity game or two, and a RAG pipeline that won the Swisscom challenge at LauzHack 2023. What stays constant is that I'd rather learn a new thing than wait for someone to hand me the right one.
Right now I'm looking for a role in or near AI and LLMs: applied ML, developer tools, research engineering, or anything adjacent where versatility beats specialization.
mocked — real Spotify API wiring is on the list
This site, some small RAG experiments over my own thesis notes, and looking for the next role.
No posts yet — my first one is cooking.
I'm based in Munich, happy anywhere in the EU, and most interested in roles near AI, LLMs, or applied ML. But, I'm also the kind of engineer who's perfectly happy writing embedded C or shipping a React app if that's what the work needs. If any of that sounds like your team, please reach out.