ML Performance Engineer

The ML Performance Engineer will oversee the entire lifecycle of ML model deployment on robots, from ML team handoff to system integration. This includes converting, optimizing, and integrating trained models for Jetson platforms using NVIDIA tools, and developing CUDA kernels and pipelines for efficient inference. Responsibilities also involve profiling and benchmarking ML workloads, identifying and resolving bottlenecks, and implementing model compression techniques for edge inference. The engineer will ensure model robustness under resource constraints, manage GPU and CPU usage on Jetson devices, build benchmarking pipelines, and collaborate with QA and systems teams for validation. Additionally, they will work closely with ML researchers to guide model architectures for edge deployability and provide technical feedback on real-time ML model feasibility within the robotics stack.