Lead · 2026
A humanoid trained for Mount Everest
Leading the first team in history to train a 29-DoF humanoid robot to summit Mount Everest (8,849 m), directing 8 engineers.
Lead Engineer at RobotEverest.
- 68/68
- 4.0 m/s
- 204.8M
Hi, I'm Adi!
I'm a high schooler interested in machine learning, robotics, and using technology to better understand the world around us.
Selected work
More work
Turned monocular trail footage into an Isaac Sim terrain for humanoid training, then compared the reconstruction against survey-grade GNSS measurements.
I built the pipeline that turns real trail footage into usable simulation terrain, including reconstruction, cleanup, benchmarking, and Isaac Sim integration. This work lays the groundwork for training robots on realistic terrain.
- 1.17 m
- 964k
- 0.5 m
Investigated how compressible a VLM’s visual-token path is, finding that 50–85% of visual tokens could be removed with less than 2% accuracy loss.
The whole study: instrumentation, five scoring methods, the allocator, and the paper draft.
- 4.6×
- 85%
- ≤2%
Built and fine tuned a YOLO-based detection pipeline to track 874 endangered great apes across 500 camera-trap videos, ready for autonomous deployment in Salonga National Park.
Led detector fine tuning, built the tracking pipeline, and ran full dataset evaluation across 500 camera trap videos.
- 0.857
- 2,257 → 301
- 874
Contact
If you find my work interesting, I’d be happy to connect. Feel free to reach out.