I'm an AI researcher at Recogna Lab (UNESP, Brazil), with a Master's in Computer Science. These days I work mostly on interpretability: I want models to be less opaque, especially in medicine, where why a model answered matters as much as what it answered. I like to think that by looking inside these machines, we also get a glimpse into our own minds.
Robotics was where I started, and I still build robots as a hobby 🤖
- Analysis of Automated Document Relevance Annotation for Information Retrieval in Oil and Gas Industry, EMNLP 2025 (Industry Track) · paper
- Class of LLMs: Benchmarking Large Language Models on the Brazilian National Medical Examination, PROPOR 2026 · paper
- Faithfulness and Uncertainty Calibration of Large Language Models in Portuguese Medical Question Answering, IEEE CBMS 2026
- The Bode Family of Large Language Models: Investigating the Frontiers of LLMs in Brazilian Portuguese, Journal of the Brazilian Computer Society, 2025 · paper
- Retrieval-Augmented Generation for Clinical Question Answering in Portuguese Drug Leaflets, PROPOR 2026 · paper
The full list is on Google Scholar.
During my undergrad I played IEEE Very Small Size robot soccer with Carrossel Caipira, UNESP Bauru's team. I also rebuilt the embedded hardware and software of the Aedromo educational robots. My bachelor's thesis applied assisted deep reinforcement learning (DQN and DDPG) to control differential-drive robots. I keep tinkering with robots in my spare time.
