Senior Machine Learning Engineer, Behavior Planning & Prediction

RESPONSIBILITIES

• Design and develop advanced machine learning models in the behavior space, specifically tailored for autonomous vehicles, utilizing deep learning and large-scale data analysis.
• Deploy scalable and efficient ML models on our autonomous vehicle platform.
• Integrate modern technologies with rigorous safety standards while maintaining cost efficiency.
• Oversee the development of new ML models end-to-end, from data strategy and initial development to optimization, production platform validation, and fine-tuning based on metrics and on-road performance.
• Lead large, multi-person projects and significantly influence the overall motion planning architecture and technical direction.
• Enable and support other engineers by coaching, leading by example, and providing high-quality code and design document reviews, as well as delivering rigorous reports from ML experiments.
• Contribute significantly to the development of essential components for end-to-end ML training and deployment, from data strategy to optimization and validation.
• Be a champion of the scientific method and critical thinking to invent state-of-the-art deep learning solutions.
• Work in a high-velocity environment and employ agile development practices.
• Collaborate closely with teams such as Perception, Simulation, Infrastructure, and Tooling to drive unified solutions.

QUALIFICATIONS

• MS or PhD in Machine Learning, Computer Science, Robotics, or related quantitative fields, or equivalent industry experience.
• 3+ years of experience with Python, PyTorch, distributed training and clean software engineering best practices (testing, code review, CI/CD).
• Comfortable writing C++ code for integration with our autonomous vehicle platform.
• 3+ years applying deep learning to sequential decision-making and prediction, including supervised/unsupervised/self-supervised learning, transfer/multi-task learning, and deep reinforcement learning.
• Extensive experience with modern sequential modeling on multi-modal sensor data (e.g., modern encoders, transformers, diffusion/latent-variable models, flow matching) and probabilistic modeling.
• 3+ years across end-to-end ML workflows: large-scale data sampling and curation, preprocessing, distributed training, ablation studies, robust evaluation, deployment, and inference optimization both for accelerators in inference time and also during training.
• Experience with learning-based decision-making for autonomous vehicles and how learned components interface between perception, motion planning and control.
• Passion for self-driving car technology and its potential to impact humanity.
• Strong communication skills with the ability to articulate concepts clearly and precisely.

NICE TO HAVES

• Published research at top-tier ML/robotics venues (e.g., NeurIPS, ICML, ICLR, CVPR, CoRL, RSS, IROS, ICRA).
• Proven track record of training and deploying large-scale multi-modal models for autonomy (e.g., behavior prediction, world models, diffusion/transformer policies) with distributed training and rigorous evaluation.
• Experience with data-centric ML: dataset design/curation at scale, active learning, auto/weak labeling, and synthetic data generation.
• Experience optimizing real-time on-vehicle inference (quantization/pruning/compilation; e.g., TensorRT, ONNX Runtime) and GPU profiling, with familiarity in production-level coding under time-limited schedules.
• Expertise across self-driving challenges (Perception, Prediction, Planning, Simulation) with emphasis on learned systems and robot motion planning.