Björn Pettersson

MILP optimisation model in Python on real 2025 operational data (tariff structures, spot prices, frequency data) evaluating FCR-D, aFRR, and spot arbitrage stacking for behind-the-meter battery assets in DK2.

Market entry and regulatory research (EU REDII, ESG, taxonomy frameworks), business case development, and VC pitch prep for an energy company, part-time alongside studies.

Co-developed PUG, a DQN reinforcement learning plugin for generative design in Rhino/Grasshopper, published on Food4Rhino. Co-authored paper presented at the Design Modelling Symposium Berlin.

BA thesis with Enyday. MILP model for FCR-D up, down, and spot arbitrage stacking for behind-the-meter energy communities in DK2. Uses real 2025 operational data.

Generative deisgn using wave function collapse for refugee camp planning. Built as a generative design tool in collaboration with CARE.

Co authored with Diego Apellániz through reserach assistant role. Reinforcement learning agent for generative spatial layout, published as a flexible reinforcement learning framework for cradle-to-cradle design.

MSc Architecture in Strategic Design & Entrepreneurship — The Royal Danish Academy & Copenhagen Business School joint programme.

During my masters I started at the CITA institute for computational design, where I fell in love with computer science through a research assistant role working on a DQN implementation that eventually got published. I switched to strategic design in my second year for a more practical and project-driven approach, but continued algorithmic design work in my master's thesis 100M — a generative design collaboration with CARE Denmark. Realising I wanted to work more theoretically, I made the switch toward a more technical career with my degree at the University of Copenhagen.