Desik Rengarajan

Papers

Total Citations

H-Index

About

Desik Rengarajan is a researcher advancing the frontiers of reinforcement learning (RL), with a primary focus on tackling the critical challenge of sparse reward environments—a pervasive obstacle in real-world applications where feedback is infrequent or binary. His most influential work, "Reinforcement Learning with Sparse Rewards using Guidance from Offline Demonstration" (2022, 26 citations), introduces a novel framework that leverages pre-collected demonstration data to guide RL agents when reward signals are insufficient, effectively bridging the gap between theoretical algorithms and practical deployment. Rengarajan further extends this paradigm in "Enhanced Meta Reinforcement Learning using Demonstrations in Sparse Reward Environments" (2022), where he integrates demonstration guidance into meta-RL to enable rapid adaptation to new tasks with minimal interaction. His recent contribution, "Federated Ensemble-Directed Offline Reinforcement Learning" (2023), addresses the emerging challenge of collaborative learning across distributed agents with limited, heterogeneous datasets—a critical step toward privacy-preserving and scalable RL systems. Through these works, Rengarajan is shaping how RL can operate under real-world constraints, making his research highly relevant for students and practitioners seeking to deploy robust, sample-efficient learning algorithms.