Max Robotics — Global Service Robot Directory + US Certification

About

Siddhant Haldar is a rising star in robot learning, whose work is redefining how robots acquire complex manipulation skills. His research centers on imitation learning, teleoperation, and policy generalization, with a focus on making robot training fast, data-efficient, and broadly applicable. Haldar’s most impactful contribution is **“Teach a Robot to FISH”** (2023, 30 citations), which introduces a versatile imitation learning framework that enables robots to learn dexterous tasks from just one minute of human demonstrations—a dramatic leap in sample efficiency. He further advanced the field with **“Watch and Match”** (2022, 8 citations), which supercharges imitation through regularized optimal transport, and **“P3-PO”** (2025), which tackles visuo-spatial generalization by incorporating prescriptive point priors. Beyond algorithms, Haldar has developed critical infrastructure: **“OPEN TEACH”** (2024) provides an open-source, user-friendly teleoperation system for robotic manipulation, while his earlier work on the **Eklavya 6.0** autonomous ground vehicle (2019) demonstrates his hands-on engineering prowess. With a growing citation footprint and a clear trajectory toward scalable, generalizable robot learning, Haldar is a researcher to watch—his work is not just advancing the science, but building the tools that will make robot learning accessible to all.