Staff, Machine Learning Engineer - BEV/Multi-Modal Perception

Job description

At Torc, we have always believed that autonomous vehicle technology will transform how we travel, move freight, and do business.

Lead BEV model development: define and execute the technical roadmap for BEV-based perception models across multiple tasks (e.g., detection, segmentation, road topology, and scene understanding).
Design advanced multi-modal architectures that fuse heterogeneous sensor data (camera, LiDAR, radar, HD maps) into unified spatial representations.
Develop foundational perception models leveraging BEV transformers, voxel-based encoders, or implicit scene representations.
Own large-scale training workflows — from data sampling strategies and augmentation pipelines to distributed training and hyperparameter optimization.
Advance model robustness and generalization, addressing long-tail conditions such as low visibility, occlusions, and rare scene configurations.
Establish evaluation frameworks for geometric accuracy, temporal stability, and cross-domain transfer performance.
Collaborate cross-functionally with sensor calibration, mapping, and fusion teams to ensure cohesive perception model interfaces.
Mentor and guide ML engineers, cultivating best practices in experimentation, code quality, and model validation.

10+ years of experience in deep learning for perception, 3D vision, and/or autonomous systems.
M.S. or Ph.D. in Computer Science, Electrical Engineering, Robotics, or related field (or equivalent practical experience).
Proven expertise in BEV modeling, 3D scene understanding, and multi-view fusion.
Strong background in multi-modal sensor fusion, particularly integrating camera and LiDAR data.
Proficiency in Python and deep learning frameworks such as PyTorch or TensorFlow.
Experience with large-scale data pipelines, distributed training, and experiment management systems.
Demonstrated leadership in driving ML model innovation and mentoring technical teams.