Malvern Madondo
Research Scientist | UChicago
Email / Linkedin / GitHub / Google Scholar
I am a Research Scientist in the Radiation and Cellular Oncology department of UChicago's Biological Sciences Division. My research focuses on developing automatic treatment planning algorithms for various radiotherapy modalities.
I obtained my 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 and Cognitive Sciences. Prior to Emory, I studied Computer Science/Information Systems and Mathematics at The College of Saint Scholastica.
Changelog
- 07/2024 — Starting as a Research Scientist at the University of Chicago in the Department of Radiation & Cellular Oncology!
- 04/2024 — Defended my PhD dissertation on "Applications of Closed-loop Control in Biomedical Interventions: From Neural Modulation to Diabetes Management" at Emory University!
- 03/2024 — Giving a lightning talk on "Learning Control Policies of Hodgkin-Huxley Neuronal Dynamics", at MCBIOS24
- 03/2024 — Poster presentation on "A Neural ODE Approach to Glycemic Control" (based on my Lilly internship) at MCBIOS24
- 11/2023 — Presented our work, "Learning Control Policies of Hodgkin-Huxley Neuronal Dynamics", at Machine Learning for Health (ML4H 2023)
Research
I am interested in applications of machine learning and mathematical optimization, particularly reinforcement learning and optimal control, in critical domains like healthcare. Some of my research efforts include developing closed-loop control solutions for neuromodulatory interventions, such as deep brain stimulation, and glycemic control systems for Type 1 Diabetes management.
Related projects include:- Closed-loop Control for Deep Brain Stimulation
[Emory University '24]Interdisciplinary collaboration with Dr. Lars Ruthotto and Dr. Nicholas Au Yong on developing adaptive control algorithms for optimizing deep brain stimulation via Reinforcement Learning and Optimal Control.
- Glycemic Control for Type 1 Diabetes Management
[IBM Research '21 | Eli Lilly '23]Implemented various control strategies for managing blood glucose levels in simulated Type 1 Diabetes patients. Leveraged RL and OC theory to model the glucoregulatory system and develop algorithms for real-time insulin delivery, reducing the risk of hyperglycemia and hypoglycemia across various meal scenarios.
- Reinforcement Learning for Sustainable Agriculture
[IBM Research '22]Developed SWATGym, a reinforcement learning environment for crop management based on the Texas A&M Soil & Water Assessment Tool(SWAT). SWATgym enables evaluation of various crop management strategies on a full growing season. It includes standard and RL-based agents for decision-making, and a reward function that captures the trade-offs between crop yield and environmental impact.
- Fair Reinforcement Learning
[IBM Research '21]Exploring bias in patient representation in healthcare and developing a fair treatment strategy by leveraging representation learning and reinforcement learning. Work done during internship with IBM Research in collaboration with the talented team at the Center for Computational Health: Mohamed Ghalwash, Zach Shahn, and Pablo Meyer Rojas.
- Reinforcement Learning and Graph Generation
[Emory University]Applying model-based Reinforcement Learning strategies in the field of graph generation, in particular, to generative models that can learn to create novel graphs efficiently.