NAAMI is launching the NAAMII Speaker Series to bring leading researchers from around the world, working at the frontiers of AI, mathematics, and computing, closer to Nepal's research community. This month's edition explores how humanoid robots learn to walk, balance, and adapt to an unpredictable world.
Join us as Niraj Pudasaini, PhD in Computer Science at CU Boulder, shares how deep reinforcement learning, whole-body humanoid control, rapid motor adaptation, and sim-to-real transfer are enabling robots to learn from experience and bridge the gap.
Between simulation and reality, adapt to unexpected forces and perform reliably in real-world environments. Drawing from his work on full-scale Unitree humanoid robots, he'll discuss rapid motor adaptation, sim-to-real transfer, and force-adaptive control for real-world deployment.
📅 Date: Friday, July 17, 2026
⏰ Time: 3:00 PM – 4:00 PM
📍 Venue: Campus France, Pulchowk, Kathmandu
🔗 Registration Link: Google Form
• Deep RL: Enables robots to learn from experience and bridge the simulation‑reality gap
• Whole‑body Control: Manages full robot dynamics for stable walking and balancing
• Rapid Motor Adaptation: Adjusts to unexpected forces in real time
• Sim‑to‑Real Transfer: Moves policies from virtual environments to physical robots
• Force‑Adaptive Control: Handles variable external loads during deployment
• Real‑World Deployment: Demonstrates full‑scale Unitree humanoid performance
Niraj Pudasaini
PhD, Computer Science, CU Boulder
BSc, Electrical Engineering, NYU Abu Dhabi
Researches deep reinforcement learning for humanoid robots, focusing on rapid motor adaptation, sim‑to‑real transfer, and force‑adaptive control.
