Humyn Labs

This is what Bengaluru’s physical AI scene looks like in one room. Last Saturday, 30+ researchers and engineers joined us to work through one question: why does egocentric hand tracking break in the wild? They challenged each other’s assumptions, disagreed openly and took the problem further together than any one of them could have alone. We’re building more rooms like this, where top minds in physical AI share their work, debate open problems and build together. If you want to be in the next one, join the waitlist. Link in the comments 👇 Ishank Gupta Manish Agarwal Sandeep Reddy Sabbella, PhD Pushpak Agrawal

100+ applications already. This Saturday, we're bringing together some very interesting folks: → An Adobe researcher who worked on egocentric hand tracking at CMU → A former CTO who built robots to help astronauts in space → A builder making mocap gloves that can sense touch → A PhD researcher at IISc's HiRO Lab An evening led by Dr. Sandeep, discussing all things egocentric hand tracking and where gaps are. Building in this space? Last few spots. We close applications in 24 hours. Sandeep Reddy Sabbella, PhD Ishank Gupta Manish Agarwal

You have a limited budget for collecting hand data for physical AI. What do you spend it on? 10,000 more clean demonstrations of pick-and-place and grasping? Or a few hundred hard sequences: fingers self-occluded or hidden behind the object, hands truncated at the frame edge, off-nominal camera viewpoints, rare grasp types? What’s the trade-off in clean vs hard in terms of performance. We would like to put on the table at our research session at Humyn Labs this Saturday. We’re bringing researchers together to examine real failure cases and debate what the next round of hand data should prioritise. 24 October · Indiranagar, Bengaluru Researchers in hand pose estimation, egocentric vision and robot learning, this one is for you! Link to apply in the comments. Sandeep Reddy Sabbella, PhD Ishank Gupta Manish Agarwal Bhoomi Arora Aachman Maheshwari

We asked Imran Nawaz what he’s working on. This is what his tabs said. Keeping Tabs EP 1. Today’s tabs: - What goes behind our validation pipeline? - Why egocentric footage gets rejected? - Why quality comes first in model-ready data and how do we protect it?

Diverse environments only help if every frame holds up. Humyn Labs delivers egocentric data that has passed multi-layer quality checks and carries dense labels, from restaurant kitchens and laundromats to construction sites, salons, and shipment hubs. Every clip is labelled for: → Depth and 6DoF pose → 21-joint hand tracking → Object detection and tracking → Dense natural-language action labels 20+ real-world environments. One QC standard.

Sight + Sound = richer annotation for Physical AI. Video shows a robot what happened. Narration tells it why. The best annotator is the person doing the task. Our egocentric videos come with live voiceover, so every label carries: → The action → The intent → The context → The world knowledge Audio-based annotation, built for Physical AI.