Tejus Gupta

Papers

Total Citations

H-Index

About

Tejus Gupta is an emerging robotics and machine learning researcher whose work sits at the intersection of reinforcement learning, imitation learning, and multi-agent systems. His research addresses fundamental challenges in enabling autonomous robots to learn complex behaviors and operate effectively in real-world environments. In his notable work on f-IRL (2020), Gupta contributed to advancing inverse reinforcement learning by developing a method that learns reward functions through expert state density matching — a significant step forward for robotic tasks where manually specifying cost functions or behaviors proves impractical. This approach offers a more flexible and generalizable framework for imitation learning, garnering 4 citations in the field. His more recent work, GUTS (2023), tackles the critical challenge of disaster response robotics, modeling search-and-rescue operations as asynchronous multi-agent active-search problems. By incorporating generalized uncertainty-aware decision-making via Thompson Sampling, Gupta's framework enables robot teams to efficiently locate targets across dangerous or expansive environments — a contribution with meaningful humanitarian implications, earning 3 citations. Though early in his research career, Gupta's work demonstrates a clear trajectory toward impactful, application-driven AI research with strong theoretical foundations in probabilistic reasoning and autonomous decision-making.