Senior ML Research Scientist, Perception Foundation Encoder

The Perception Foundation Encoder is the first processing stage, converting sensory streams (cameras, LiDAR, radar, and audio) into a unified feature representation. You will design and train scalable multi-modal models, exploring distillation and compression to balance accuracy and onboard latency. Collaborate with cross-functional teams to advance object detection, tracking, online mapping, and scene understanding, while maintaining code quality and robustness in production. You will leverage Ph.D.-level research experience in multi-modal or large-scale models and 3+ years of ML project work to push the state of the art. Strong software engineering in Python and C++ is required, along with a track record of publications at major conferences. This role offers the opportunity to impact perception, decision-making, and learned behavior modules across the stack.