6 months ago
Remote, United States or Seattle, WA, USASenior
Responsibilities
- Architect and advance perception algorithms for scene understanding, obstacle detection, terrain modeling, sensor fusion, and machine awareness.
- Develop, train, evaluate, and deploy machine learning models for object detection, segmentation, depth estimation, and terrain understanding.
- Optimize perception models for real-time inference on edge compute platforms.
- Build scalable data pipelines for training, evaluation, and model iteration.
- Integrate perception outputs with robotics, controls, localization, SLAM, mapping, planning, and safety systems.
- Develop robust, testable, and observable perception interfaces and debug systems using field data.
- Research and apply state-of-the-art machine learning and perception methods on autonomous machines deployed in the field.
Requirements
- Experience designing, training, and evaluating machine learning models and shipping production models.
- Strong understanding of ML optimization, gradients, regularization, bias-variance tradeoffs, training dynamics, and data imbalance.
- Experience with 2D and 3D machine perception, computer vision, sensor calibration, classification, detection, segmentation, and tracking.
- Understanding of image representations, augmentations, preprocessing, CNNs, transformers, evaluation metrics, and precision/recall, mAP, and IoU tradeoffs.
- Proven track record of deployed ML-based perception systems, ideally for autonomous vehicles.
- Excellent software development skills in Python and proficiency in TensorFlow, PyTorch, and C++.
- Ability to solve complex problems creatively and work hands-on with real-world hardware and field deployments.
Benefits
- Company-funded medical, dental, vision, 401k, and life insurance.
- Gourmet food and other company perks.
- Strong onsite collaboration at AIM offices, labs, and proving grounds in the east side of the Greater Seattle area.
- Opportunities to travel to sites across the Americas, Australia, Africa, and other regions.
Categories
ML EngineeringRobotics
