Summary
Steven Abreu is an AI researcher and engineer with nine years of experience, currently building self-improving AI agents at MakerMaker.AI in San Francisco. He holds a PhD focused on novel compute paradigms, hardware for AI, interpretability, safety, and recurrent neural networks, and has interned or researched at Intel, Google, ETH Zürich, and Ghent on neuromorphic, photonic, and low-precision ML systems. His work spans from practical production ML (real-time pricing at Bloomberg) to brain-inspired and mixed-signal computing, giving him a rare fluency across algorithms, hardware accelerators, and evaluation for safety and scalability. Notably, he has contributed to scaling laws and non‑transformer LLM approaches on new accelerators, reflecting a blend of theoretical depth and hands-on systems engineering.
9 years of coding experience
6 years of employment as a software developer
Bachelor of Science (incomplete) Physics, Bachelor of Science (incomplete) Physics at Karlsruhe Institute of Technology (KIT)
Bachelor of Science (B.Sc.) Computer Science, Bachelor of Science (B.Sc.) Computer Science at Jacobs University Bremen
Doctor of Philosophy - PhD Artificial Intelligence, Doctor of Philosophy - PhD Artificial Intelligence at University of Groningen
Bachelor of Science (B.Sc.) Computer Science, Bachelor of Science (B.Sc.) Computer Science at Carnegie Mellon University
German, English, Spanish, Dutch, Portuguese