BUILT IN PUNE. DEPLOYED EVERYWHERE.
PanditaAI is an operating unit of Shailka Systems Pvt. Ltd., headquartered in Pune, India. We are a father-and-son founding team building the open marketplace of physical skills for humanoid robots.
THE FOUNDING STORY
Sumit Agrawal — 8+ years across AI, Web3, and cloud-native systems, operator of the MCP/WebMCP tooling ecosystem with over $1.2B AUM across his portfolio — teamed up with his father, Shailendra Agrawal, an electrician and mechanical specialist with a lifelong fascination for humanoid hardware.
The observation: humanoid robots are shipping, but they arrive without skills. Every OEM is re-training from scratch. Meanwhile, human experts — chefs, martial artists, gardeners — hold expertise that can only be preserved if someone captures it. PanditaAI is that bridge.
Core Directives
OPERATIONAL STANDARDS 2024.v1SUMIT AGRAWAL · CEO
8+ years across AI, Web3, and cloud-native infrastructure. Operator of the MCP/WebMCP tooling ecosystem with over $1.2B AUM across his portfolio. Leads product, engineering, and capital strategy.
THE MISSION
Capture human expertise before it disappears, and make it downloadable to any NVIDIA-compatible humanoid. An open skill layer for the physical economy.
SHAILENDRA AGRAWAL · PRESIDENT
Electrician, mechanic, and hardware specialist. Lifelong passion for humanoid robotics. Owns the physical build, hardware integration, and motion-capture studio operations.
System FAQ
QUERY_PORT_8080What constitutes a “Physical Skill” in the PanditaAI framework?
addA physical skill is a TensorRT-optimized policy trained in NVIDIA Isaac Sim from expert human demonstration. Sizes range from 500MB to 2GB. Once licensed, a skill deploys OTA to a robot's Jetson Orin / IGX compute and runs at sub-5ms inference.
Which hardware platforms are currently supported?
addOur deployment target is any humanoid running NVIDIA Jetson Orin or IGX. We're launching with Unitree-class humanoids as the reference platform and expanding to additional OEMs as pilots ship.
How does PanditaAI handle edge-case safety?
addEvery skill is validated against randomized physics and friction parameters in Isaac Sim before publishing. Operators can enforce torque, speed, and workspace limits at runtime on Jetson / IGX through the skill manifest.
Establish Contact
Secure transmission line for partnerships and technical inquiries.