Nikhil Sardana

About

Nikhil Sardana is a researcher advancing the frontier of reinforcement learning (RL) for real-world, embodied systems. His work focuses on bridging the gap between simulated RL benchmarks and the complexities of physical agents operating in dynamic, non-stationary environments. Sardana’s major contribution is the formalization of **autonomous reinforcement learning**, a paradigm that addresses the critical challenge of agents learning continuously from their own stream of experience—without human resets or engineered reward functions. His highly cited 2021 paper, *“Autonomous Reinforcement Learning: Formalism and Benchmarking,”* provides the foundational framework and evaluation suite for this emerging subfield, enabling researchers to systematically study how agents can acquire and refine skills through open-ended interaction. This work has garnered 3 citations and is shaping how the community thinks about deploying RL in robotics and other real-world applications where constant human oversight is impractical. Sardana’s research is essential reading for anyone interested in making RL truly autonomous and deployable in the wild.

Research Focus

Key Achievements

Top Papers

1. 1
   Autonomous Reinforcement Learning: Formalism and Benchmarking

   3 citations · 2021

Key Collaborators

Contact & Links