Research Scientist at the University of Chicago · Department of Radiation & Cellular Oncology
I develop AI-driven treatment planning algorithms at the intersection of deep reinforcement learning, optimal control, and clinical oncology. My current work focuses on automatic treatment planning for radiotherapy modalities using patient-specific RL frameworks.
I hold a PhD in Computer Science from Emory University, where I was advised by Dr. Lars Ruthotto and supported by a Google PhD Fellowship in Computational Neural & Cognitive Sciences.
- AI for Radiation Therapy — deep RL for proton therapy treatment planning (head-and-neck cancer)
- Closed-loop Biomedical Control — neuromodulation via Hodgkin-Huxley dynamics, glycemic control via Bergman minimal model
- Optimal Control + ML — neural ODEs, Pontryagin's Maximum Principle, Hamilton-Jacobi-Bellman equations
- Fairness in Healthcare AI — equitable treatment strategies across diverse patient populations
| Repository | Description |
|---|---|
| protonRL | RL framework for automatic treatment planning in proton radiotherapy |
| HH-control | Optimal control & ML for neuromodulation based on Hodgkin-Huxley dynamics |
| SWATGym | IBM Research open-source RL environment for crop management |
- M. Madondo et al. "Patient-Specific Deep RL for Automatic Replanning in Head-and-Neck Cancer Proton Therapy" — MLHC 2025 · arXiv
- M. Madondo et al. "ProtonRL: Patient-Specific RL for Automatic Replanning in Proton Therapy" — AAPM 2025 · Therapy Physics Scientific Session: AI in the Clinic
- M. Madondo et al. "Learning Control Policies of Hodgkin-Huxley Neuronal Dynamics" — ML4H 2023 · arXiv
- M. Madondo et al. "A SWAT-based RL Framework for Crop Management" — AISG @ AAAI 2023 · arXiv
Full list → Google Scholar