MIT
Graph-aware learning
Graph representations and message passing for learning in coupled systems.
Read moreMachine learning research
Hi, I'm Raphael.
Machine learning researcher
in Cambridge, Massachusetts.
At MIT, I study graph representations and scalable multi-agent learning. I also collaborate with researchers at Harvard on social foraging, LLM interpretability, and language bias.
On my mind
ML theory, statistics & probability. Lately, post-training, multimodal models, and AI safety.
Selected research
All researchMIT
Scalable multi-agent systems
Coordination and action selection for scalable multi-agent reinforcement learning.
Read moreHarvard · Collaboration
Social foraging & agent behavior
Collaborative research on recurrent agents, social foraging, and behavioral analysis.
Read moreHarvard · Collaboration
LLM interpretability & language bias
Evaluating how language models preserve or shift meaning across languages.
Read moreSelected projects
All projectsI also enjoy building products and exploring what it takes to turn an idea into something people use.
AMBR · AI for Bio
Machine learning for antibiotic-resistance research. Built with the team that won Hack-Nation 2026.
PrognoseWerk
Probabilistic demand forecasting, with research into Bayesian deep learning and calibrated uncertainty.
Language-guided 3D editing
Grounding language in a scene graph, validating structured actions, and executing edits in NVIDIA Isaac Sim.
Nyra
A personal AI assistant that follows a conversation and offers help when it is useful. Starting with voice, timing, and context.
Along the way
A few highlights.
Work, teams and opportunities
I'm grateful to have been part of.
- Hack-Nation Global AI HackathonGlobal winner with the AMBR team2026
- IDEA League Challenge ProgrammeSelected participant with travel scholarship2025–2026
- RWTH Dean's ListBachelor's and master's studies · Top 5% in the awarded cohorts2024–2025
