About

I studied physics: a bachelor’s in Madrid, then a master’s in nuclear, particle and astrophysics (KTA) at TUM in Munich. Since then I’ve worked on data and trading systems, first in research at the FRM II neutron source, now as a senior quantitative developer at Galleon Ventures, always in Python: investments, backtesting engines, trade execution and data pipelines, with Svelte for the interfaces.

These days I build with whole teams of AI coding agents working in parallel. What makes that work is knowing the code and the infrastructure underneath in depth: I split the work, give each agent the right context, and review everything that comes back, so I deliver many times more without lowering the bar on quality. FreeDeskAI grew out of that way of working.

On the side I co-authored PEP 718, a Python Enhancement Proposal for subscriptable functions, sponsored by Guido van Rossum, the creator of Python, published a paper at Computers and Games 2024 on the maths of the sorting game PlaceIt, write Typst libraries, answer Python and pandas questions on Stack Overflow, play chess, and shoot 35mm film.