01 · Adaptive control
Robust RL for time-varying systems
Hybrid actor-critic and bounded extremum-seeking controllers that preserve fast learned behavior while adapting to drift and nonstationary dynamics.
Read the ACC 2026 paperReinforcement learning & control · Los Alamos, New Mexico
Shaifalee Saxena
Ph.D. Student & Graduate Research Assistant
I develop robust learning-based control methods for robotics, particle accelerator tuning, and autonomous systems operating under distribution shift.
Best Paper Award · IROS AIM-Ctrl 2026
About
Learning systems that remain useful when the world changes
I am a Graduate Research Assistant in the Adaptive Machine Learning group at Los Alamos National Laboratory and a Ph.D. student in Electrical and Computer Engineering at the University of New Mexico.
At LANL, I work under the mentorship of Dr. Alexander Scheinker. At UNM, I am advised by Dr. Rafael Fierro through the MARHES Lab and the Agile Manufacturing Lab.
My research combines reinforcement learning, extremum seeking, robotics, representation learning, and adaptive control. I am particularly interested in controllers that can recognize distribution shift, stop trusting an unreliable learned policy, and adapt safely online.
Positions
Graduate Research AssistantLos Alamos National Laboratory · 2025–present
Research AssistantUniversity of New Mexico · 2024–present
Senior Flight Controls EngineerAirbus · 2021–2024
Senior Control Systems EngineerTata Aerospace and Defense · 2018–2021
Education
Ph.D., Electrical & Computer EngineeringUniversity of New Mexico · 2024–present
B.Tech., Aerospace EngineeringIIST · 2014–2018
Research
Robust intelligence for dynamic physical systems
My work connects learning, adaptation, and control across accelerator and robotic platforms.
02 · OOD detection
Learning when to trust RL
Mahalanobis-guided latent OOD detection for deciding when a controller should stop using a learned RL policy and switch to robust bounded extremum seeking.
Read the ICML RLxF paper03 · Autonomous systems
Visual control and space robotics
Compact visual representations, failure detection, and corrective control for autonomous systems, including satellite solar-panel deployment.
Explore all research projectsNews
Recent updates
-
Received a Best Paper Award at the IROS AIM-Ctrl Workshop 2026 for our work on reinforcement learning and adaptive control for robotic manipulation. Paper
-
Gave an invited talk at CAARI titled "Bridging Deep Reinforcement Learning and Adaptive Feedback for Time-Varying Systems." CAARI
-
Our paper was accepted at the IROS AIM-Ctrl Workshop 2026. Paper
-
Received the LANL Best Student Award.
-
Received SPOT Awards from the AOT-IC Group at Los Alamos National Laboratory.
-
The ICML RLxF Workshop 2026 poster was presented by my mentor, Dr. Alexander Scheinker, as I was unable to attend in person. View poster
-
Attended the DL4Sci Summer School. Summer school
-
Presented a poster at the CVPR CV4Science Workshop 2026. View poster
-
Our LANL Innovative Research (IR) proposal, "Agentic Adaptive Robotics for Hazardous Maintenance & Glovebox/Workcell Operations," was accepted.
-
Our paper was accepted at the ICML RLxF Workshop 2026. Paper
-
Our abstract was accepted at the CVPR CV4Science Workshop 2026.
-
Serving as a reviewer for IROS 2026.
-
Received an ACC travel grant to present my paper at ACC 2026. Paper
-
Serving as a reviewer for CDC 2026.
-
Accepted to the DL4Sci Summer School. Program
Selected work
Recent publications
Learning-based control methods designed for changing, uncertain environments.
ACC 2026
Improved Robustness of Deep Reinforcement Learning for Control of Time-Varying Systems by Bounded Extremum Seeking
Community
Service & recognition
Peer review
Conference reviewer
Reviewer for the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2026) and the IEEE Conference on Decision and Control (CDC 2026).
Recognition
Awards and support
Best Paper Award at IROS AIM-Ctrl 2026, LANL Best Student Award, AOT-IC SPOT Awards, ACC 2026 travel grant, Airbus SPOT Award, Tata Best Performer and Team Award, and ICONS 2018 second-best-paper award.
Engagement
Invited talks and programs
Invited speaker at CAARI 2026 for “Bridging Deep Reinforcement Learning and Adaptive Feedback for Time-Varying Systems,” and participant in the 2026 Deep Learning for Science Summer School.