About this role
ABOUT KOVARI
Kovari develops frontier technology to transform physical businesses. They partner with established operators to change how they run, and build and operate businesses of their own through their subsidiaries.
Kovari is a small team of builders, operators, and researchers. They deploy on-site alongside the businesses they serve and iterate in the field. The company has raised over $6M from leading investors and is based in San Francisco.
THE OPPORTUNITY
You will research and ship manipulation policies that work reliably in the field. You own the loop from data to model to deployment, and you see your policies run on robots in live operations.
WHAT YOU'LL DO
- Research and develop high-reliability manipulation policies built for fast deployment and iteration
- Operate in a fast data flywheel across multiple modalities
- Deep debug failure modes in transformer and diffusion policies deployed in the field
- Optimize policies for real-time inference on edge hardware
YOU SHOULD HAVE
- Experience deploying robot policies on hardware, including model-based, RL, or imitation learning approaches
- Sim-to-real or real robot data experience
- Experience with multimodal inputs including vision, depth, force/torque, and proprioception
- Experience with edge inference optimization using TensorRT, ONNX Runtime, or TVM
- Experience with CUDA kernel optimization
- A bias toward action: train fast, deploy early, learn from failure
NICE TO HAVE
- Publications at CoRL, RSS, ICRA, NeurIPS, or similar venues