Senior, ML Engineer - Offline Perception
Core
Design, implement, test, and deploy offline object detection, tracking, and fusion modules to automatically create annotations from logged sensor data (Cameras, Lidars, Radars) for autonomous truck perception.
Role type
Senior, hands-on IC machine-learning engineer (offline perception & pseudo-labeling)
Builds
High-quality sensor data annotations (2D/3D bounding boxes, trajectories, segmentations) used to train perception models and generate simulation data.
Domain
Autonomous driving / Trucking / Computer Vision
Deliverable
production ML models
Required skills
Active Learning, Pseudo-labeling, Computer Vision, Deep Learning, 2D/3D Object Detection, Tracking, Sensor Fusion, Semantic Segmentation, Scaled MLOps, Distributed machine learning frameworks, Model Data Curation, Python software development
Preferred skills
SLAM, BEV, ML Frameworks, experiment tracking, model registry, MLFLow, Weights and Biases, Parquet data processing, VDI, CI Systems, Docker, AWS storage and processing infra, vector databases, Data Visualization, Cloud Development, Cloud-based orchestration, Model Inference Orchestration
Technologies
PyTorch, Lightning, Ray, Parquet, PyArrow, Daft, Pandas, GitHub Actions, Docker, AWS (S3, ECS, Lambda, Dynamo, Step Functions, Athena), Terraform, OpenGL, 3.js, foxglove, Tableau, MCAP
Responsibilities
Design and deploy offline perception models/algorithms; define and implement data ingestion, preparation, curation, and governance; measure and track auto labeling quality; develop guidelines and standards for ML model deployment; provide technical guidance, coaching, and mentoring to team members.
Seniority
Senior, hands-on IC with project leadership