Machine Learning Engineer, Perception

Description

The company builds AI that lets robots adapt, learn and work in real environments, aiming to close the skilled labor gap across industries.

Its systems combine perception, reasoning and control, and are built to be reliable, aware of risk and better over time as robots run in the field.

The team works where welding science meets artificial intelligence, on hard problems alongside other experienced engineers.

Responsibilities

• Experienced: Develop and test machine learning methods for weld perception, such as point cloud registration, seam detection and joint geometry estimation, extending them to more joint types and part shapes over time.
• Experienced: Build and maintain pipelines that ingest annotated 3D scans, generate synthetic data and manage datasets for repeated model improvement.
• Experienced: Run careful evaluations that analyze failure modes, false positive and false negative rates, and registration accuracy against set thresholds, then report results so next steps are clear.
• Experienced: Put trained models into production ROS robotics services with low-latency inference that works with deployed cell setups.
• Experienced: Write clean, tested Python, take part in code and experiment reviews, and document methods, parameters and results.
• Senior/Staff: Lead research, development and production rollout of advanced perception algorithms for point cloud registration, seam detection and real-time in-process tracking using structured light, RGB and stereo sensors.
• Senior/Staff: Design and lead experiments on modern deep learning models, including transformer and geometric feature learning architectures.
• Senior/Staff: Design and lead real-time perception systems, such as seam tracking during welding, that fuse sensor data with probabilistic state estimation like Kalman filtering for reliable weld results.
• Senior/Staff: Own the full ML lifecycle, from dataset design and annotation strategy through training, benchmarking and fleet deployment, using clear go/no-go evaluation criteria.
• Senior/Staff: Set up distributed training and hyperparameter optimization workflows, and set direction on data acquisition, annotation tooling and the use of synthetic versus real scan data.
• Senior/Staff: Mentor engineers at all levels, giving technical guidance on perception systems and ML methods.

Requirements

• Master's or Ph.D. in computer science, robotics or a related field such as computer vision, ML or perception. A Bachelor's with strong production ML experience is also considered.
• 3+ years (Experienced) or 7+ years (Senior/Staff) in real-world robotics or industrial ML.
• Strong Python skills and hands-on PyTorch work, including training, evaluating and deploying deep learning models in production.
• Experience with 3D point cloud data and libraries such as Open3D, including surface segmentation, spatial queries and point-wise labeling.
• Familiarity with 3D deep learning architectures such as PointNet++, GeoTransformer, or similar transformer or graph-based methods for geometric data.
• Able to integrate ML models into production robotics services within ROS-based architectures and containerized deployment environments.
• Senior/Staff: A proven record leading end-to-end ML projects from dataset design through fleet deployment with rigorous go/no-go frameworks, plus experience architecting distributed training and hyperparameter optimization workflows.

Conditions

Openings are available at L3/L4 (experienced), L5 (senior) and L6 (staff) levels.

Benefits

• Free lunch every day.
• Flexible paid time off, taken when you need it.
• Medical, dental and vision coverage.
• Fully paid parental leave of 6 weeks, plus an additional 6, 8 weeks for birthing parents, for 12, 14 weeks in total.
• 401(k) retirement plan through Empower.
• Referral bonuses for bringing new people onto the team.