Research Scientist · University of Chicago
Developing AI-driven treatment planning algorithms at the intersection of deep reinforcement learning, optimal control, and clinical oncology.
I am a Research Scientist in the Biological Sciences Division at the University of Chicago, working in the Department of Radiation & Cellular Oncology. My research focuses on developing automatic treatment planning algorithms for various radiotherapy modalities using deep reinforcement learning and mathematical optimization.
I obtained my PhD in Computer Science from Emory University, advised by Dr. Lars Ruthotto and supported by a Google PhD Fellowship in Computational Neural and Cognitive Sciences. My doctoral research developed closed-loop control solutions for neuromodulatory interventions and glycemic control systems. Prior to Emory, I studied Computer Science and Mathematics at The College of St. Scholastica.
University of Chicago · Dept. of Radiation & Cellular Oncology, Tian Lab
Developing deep RL algorithms to optimize proton therapy treatment plans for head-and-neck cancer. Designing scalable simulated environments in Python and PyTorch to accelerate prototype-to-clinic translation of AI-driven treatment planning solutions across multiple modalities.
Emory University · Machine Learning & Inverse Problems Group
Developed closed-loop control frameworks for DBS neuromodulation (Hodgkin-Huxley dynamics) and Type 1 diabetes management (Bergman minimal model), integrating neural networks with Pontryagin's Maximum Principle and HJB equations to enable real-time adaptive treatment in silico. Secured over $350K in competitive research funding including the Google PhD Fellowship.
Eli Lilly & Company · Advanced Analytics & Data Science
Developed a hybrid control framework combining neural ODEs with PID control and HJB-based optimal control for automated insulin delivery. Modeled glucose-insulin dynamics via the Bergman minimal model, designed reward metrics and adaptive dosing strategies for real-time glucose regulation.
IBM Research · Responsible & Inclusive Technologies
Created and open-sourced SWATGym for precision agriculture (adopted by 5+ research groups). Developed fairness-constrained RL algorithms for equitable diabetes management achieving >87% Time-In-Range across diverse patient populations.
Meta · Crisis Response & Facebook University
Built crisis assistance tools boosting Community Help platform engagement by 39%. Co-designed an iOS app with location-based features and image recognition via IBM Watson API.
I am interested in applications of machine learning and mathematical optimization, particularly reinforcement learning and optimal control, in critical domains like healthcare. My current focus is on AI-driven treatment planning for radiation therapy, with prior work spanning neuromodulation, glycemic control, sustainable agriculture, and algorithmic fairness.
University of Chicago · 2024-Present
Developing patient-specific deep RL frameworks (ProtonRL) for automatic replanning in head-and-neck cancer proton therapy, improving clinical decision support and accelerating treatment workflows.
Emory University · 2019-2024
Adaptive control algorithms for optimizing deep brain stimulation via RL and optimal control, in collaboration with Dr. Lars Ruthotto and Dr. Nicholas Au Yong.
IBM Research · Eli Lilly · 2021-2023
Developed control strategies leveraging RL and OC theory for real-time insulin delivery, reducing hyperglycemia and hypoglycemia risk across diverse patient populations and meal scenarios.
IBM Research · 2022
Created SWATGym, an RL environment for crop management based on the Texas A&M SWAT model, enabling evaluation of management strategies that balance crop yield with environmental impact.
IBM Research · 2021
Explored bias in patient representation in healthcare and developed fair treatment strategies leveraging representation learning and RL, in collaboration with the Center for Computational Health.
Full list on Google Scholar. * denotes equal contribution. Surname: Madondo.
Patient-Specific Deep Reinforcement Learning for Automatic Replanning in Head-and-Neck Cancer Proton Therapy
Machine Learning for Health Conference (MLHC 2025)
ProtonRL: Patient-Specific Deep Reinforcement Learning Framework for Automatic Replanning in Proton Therapy for Head-and-Neck Cancer
AAPM 2025 — Therapy Physics Scientific Session: AI in the Clinic
Deep learning-based applicator selection between Syed and T&O in high-dose-rate brachytherapy for locally advanced cervical cancer: a retrospective study
Physics in Medicine and Biology (2025)
Automated Treatment Planning for Interstitial HDR Brachytherapy for Locally Advanced Cervical Cancer using Deep Reinforcement Learning
arXiv:2506.11957 (2025)
Forecasting land-based environmental variables using similarity analysis and temporal graph convoluational neural networks
US Patent US-20240112442-A1
Learning Control Policies of Hodgkin-Huxley Neuronal Dynamics
Machine Learning for Health (ML4H) Findings Track, 2023
A SWAT-based Reinforcement Learning Framework for Crop Management
AI for Social Good Workshop at AAAI (2023)
Forecasting Soil Moisture Using Domain Inspired Temporal Graph Convolution Neural Networks
IJCAI-23 Special Track on AI for Good, pp. 5897-5905