Marco Selvi is a Research Scientist in the AI Research team within AI & Robotics at EIT.
He previously worked at Meta on RL fine tuning for Code World Models, and at DeepMind on RLEF and RLHF for code agents, as well as building standalone RL training regimes and environments for code automation and discovery. He was one of the core authors of AlphaDev (published in Nature in 2022). Before DeepMind Marco worked at Jukedeck on automated symbolic music generation.
Marco is a graduate of the University of Cambridge, where he researched and read Physics and Computer Science. He holds an MPhil in Scientific Computing, an MSci in Physics and an MA in Natural Sciences. His masters鈥 research focused on 3D cellular automata simulations and relativistic gravitational dynamics.
Outside of work Marco enjoys playing the violin and piano, reading assiduously, traveling, and playing board and video games.
鈥
Awards and certifications
Publications
We鈥檙e gathering the greatest minds.
EIT is growing rapidly and we are recruiting at pace. Thrive in a dynamic and fast-paced work environment, learning and growing every day alongside experts in science and technology.