Machine Learning Engineer: Perception
Bedrock Robotics Inc5 months ago
Responsibilities
- Design and train state-of-the-art early-fusion models that combine raw LiDAR and camera data for object detection and semantic segmentation.
- Build perception systems robust to dynamic occlusion, dust, snow, rain, and high-vibration conditions.
- Optimize models for inference on embedded hardware and debug sensor-calibration drift, latency bottlenecks, and other system-level issues.
- Collaborate with other teams to create state-of-the-art representations for downstream use cases.
- Manage or build data pipelines, assess data alignment and ground-truth quality, and evaluate perception-system corner cases.
Requirements
- 3+ years of experience taking deep-learning models from research into real-world production using PyTorch, TensorFlow, or JAX.
- Deep understanding of SE(3) transformations, homogeneous coordinates, and intrinsic and extrinsic sensor calibration.
- Practical experience with feature-level early-fusion architectures such as BEVFusion, TransFuser, or PointPainting.
- Experience with modern transformer-based object-detection architectures including DETR and PETR, plus temporal models such as PETRv2 and StreamPETR.
- Expertise in Python and ability to read and write systems code in C++ or Rust, including understanding of memory management and real-time constraints.
- Understanding of robotics data infrastructure, ground-truth quality, model evaluation, corner cases, and statistical properties of perception systems.
- Experience with computer vision and/or LiDAR-based perception systems.
- Preferred experience with occupancy grids, NeRFs, or voxel-based representations for terrain mapping.
- Preferred publications in conferences such as ICRA, IROS, CVPR, ECCV, ICCV, CoRL, or RSS.
About Bedrock Robotics Inc
Bedrock Robotics builds autonomous control systems that retrofit heavy construction equipment, enabling driverless operation on large infrastructure and industrial projects. The San Francisco–based, privately held company, founded in 2024, deploys its technology with contractors and project owners to speed schedules and improve job-site safety. Its business centers on upgrading existing fleets and operating them in the field, with software, sensors, and integration services tailored to construction workflows.