Machine Learning Engineer: Perception

The role

You will design and train state-of-the-art models that fuse Lidar and Camera data for object detection and semantic segmentation. You will build perception systems robust to dust, snow, rain, and vibration, and optimize models for inference on embedded hardware.

What you will do

• Design Early Fusion Architectures by developing and training models like BEV-based transformers.
• Build perception systems to handle dynamic occlusion, particulates, and high-vibration conditions.
• Optimize models for inference on embedded hardware.
• Debug system-level issues such as sensor calibration drift and latency bottlenecks.
• Collaborate with other teams to create state-of-the-art representations for downstream use cases.

Required qualifications

• 5+ years of experience taking deep learning models from research to real-world production using PyTorch, Tensorflow, or JAX.
• Deep understanding of SE(3) transformations, homogeneous coordinates, and intrinsic/extrinsic sensor calibration.
• Ability to project a 3D Lidar point onto a 2D image pixel accurately.
• Practical experience with architectures that fuse modalities at the feature level.
• SOTA Object Detection experience with modern transformer-based architectures including similar temporal models.
• Expert in Python.
• Comfortable reading and writing systems code in C++ or Rust.
• Understanding of memory management and real-time constraints.
• Willingness to dig into data infrastructure to ensure ground truth quality.

Nice to have

• Experience working with occupancy grids, NeRFs, or voxel-based representations for terrain mapping.
• Published work in conferences such as ICRA, IROS, CVPR, ECCV, ICCV, CoRL, or RSS

About the company

Bedrock retrofits heavy equipment like excavators with autonomy for construction job sites. The team includes veterans from Waymo, Segment, and Uber Freight.