Agustin Castellano
PhD Candidate , Electrical and Computer Engineering, Johns Hopkins University
Barton Hall, 320B
3400 N Charles St
Baltimore, MD, USA
Hi there! I’m Agustin Castellano, a 5th year PhD student at Johns Hopkins University advised by prof. Enrique Mallada. I mostly work on Reinforcement Learning algorithms for safety-critical systems, doing a healthy mix of theory and applications.
Prior to starting at Hopkins, I got the M.Sc. degree in Electrical Engineering from Universidad de la República, Uruguay, under the guidance of prof. Juan Bazerque.
I’m building this webpage as we speak, so stay tuned!
news
| Jan 27, 2026 | Happy to share that we just had a paper accepted with my dear friend and (now) co-author Tomás Tapia. Our work on Reachability Guarantees on Energy Storage will be presented at the 2026 Power Systems Computation Conference. |
|---|---|
| Jan 22, 2026 | One paper accepted to L4DC! Our work on “Data-driven Acceleration of MPC with Guarantees” has been accepted as an oral presentation for the Learning for Dynamics and Control Conference, to be held in University of Southern California in Mid June this year. |
| Jun 09, 2025 | I will be attending the 2nd annual Reinforcement Learning conference, where I will present our most-recent work “Non-parametric Policy Improvement in Continuous Action Spaces via Expert Demonstrations”. |
| Apr 01, 2025 | I was awarded the Johns Hopkins Whiting School of Engineering Excellence in Teaching Award for my role as lead Teaching Assistant in “Foundations of Reinforcement Learning”. |
| Mar 25, 2025 | I’ll be interning at Los Alamos National Laboratory this Summer, working under Harsha Nagarajan on Reinforcement Learning-accelerated solutions to non-convex optimization problems! |
selected publications
-
Learning to act safely with limited exposure and almost sure certaintyIn Transactions on Automatic Control (TAC) , May 2023 -
Reinforcement Learning with Almost Sure ConstraintsIn Learning for Dynamics and Control , May 2022