Senior ML Engineer – Autonomous Mapping & Perception

Socket.dev; Ann Arbor (MI); Hybrid; USD 177,000 - 213,000; Full time

Benefits offered by this job

Bonus and stock options

Fully paid medical/dental/vision

401K with 6% match

Flexible vacation

Holiday closures

AD+D and Life Insurance

Job summary

Torc is seeking a highly skilled engineer to advance offline perception, mapping, and 3D reconstruction for autonomous driving. You will lead lane-line detection pipelines, develop pose estimation for accurate localization, and guide a team through ML-driven annotation workflows.

The role emphasizes scalable cloud-based development, strong Python, and distributed ML frameworks. The position offers a competitive package including bonus and stock options, full medical coverage, a 401K match, and

Qualifications

• Experience in lane line annotation creation or automatic mapping.
• Familiarity with lane line detection ML models.
• Familiarity with pose estimation.
• Active Learning & Pseudo-labeling – CV, DL, model training.
• Two of: Object Detection, Tracking, Sensor Fusion, Semantic Segmentation, SLAM, BEV.
• ML ops and tooling: MLflow, Weights & Biases, experiment tracking.
• Distributed ML with PyTorch, Lightning, Ray.
• Data curation with Parquet (PyArrow, Pandas).
• Python development; cloud-based environments; CI systems; Docker.

Responsibilities

• Design, implement, test and deploy offline 3D reconstruction and lane-line detection modules to generate annotations.
• Build and refine lane line annotation pipelines for automated map creation.
• Develop pose estimation algorithms for accurate localization and sensor fusion.
• Provide project leadership and guide junior engineers across project execution.
• Stay updated with AI/ML advances in autonomous driving and mapping.
• Develop offline perception and mapping models with disciplined software processes.
• Define data ingestion, preparation, curation and governance for large datasets.
• Assess capabilities and propose improvements aligned with strategy.
• Measure auto-labeling and map quality to meet requirements.
• Create information products to support visualization and data access.
• Evaluate technical advances to boost productivity and quality.
• Develop guidelines for analytics and ML model deployment.
• Provide technical guidance and mentoring to team members.

Skills

Python

Pose estimation

Sensor fusion

Lane line annotation

Object detection

SLAM

MLOps

PyTorch

Ray

Docker

Education

Bachelor's degree in CS/Robotics/EE

Master's degree in CS/Robotics/EE

Tools

Docker

GitHub Actions

CI/CD

Cloud development environments

Job description

Torc is seeking a highly skilled engineer to advance offline perception, mapping, and 3D reconstruction for autonomous driving. You will lead lane-line detection pipelines, develop pose estimation for accurate localization, and guide a team through ML-driven annotation workflows.

The role emphasizes scalable cloud-based development, strong Python, and distributed ML frameworks. The position offers a competitive package including bonus and stock options, full medical coverage, a 401K match, and