Staff ML Engineer: Learned Emergency Trajectory for AVs

EngineersOfAI; Sunnyvale (CA); Hybrid; USD 180,000 - 260,000; Full time

Job summary

Wayve is seeking a Staff Machine Learning Engineer to lead the technical direction of a learned emergency trajectory model for low-frequency, high-consequence maneuvers within the AV Core team. You will drive problem definition, modelling, evaluation, integration, and deployment evidence across teams and partners.

Responsibilities include setting strategy, training policy models, and building data strategies for rare events, while ensuring robust verification and safe integration with the

Qualifications

• Track record of staff-level leadership guiding ambiguity in ML programmes.
• Experience developing learned trajectory‑generation or policy models for embodied systems.
• Hands-on with behaviour cloning, reinforcement learning, or related methods.
• Strong Python and PyTorch ML engineering skills with scalable training/evaluation systems.

Responsibilities

• Set the technical strategy and roadmap for the emergency trajectory model with measurable criteria.
• Design and train trajectory-generating policies using suitable ML methods.
• Build a data strategy for rare emergency cases including simulation and augmentation.
• Create rigorous open-loop and closed-loop tests for safety-critical maneuvers.
• Lead integration into the shared driving stack and communicate risks and trade-offs.

Skills

Staff-level leadership

Learned trajectory generation

Behaviour cloning / RL

Python

ML engineering

Tools

PyTorch

Job description

Wayve is seeking a Staff Machine Learning Engineer to lead the technical direction of a learned emergency trajectory model for low-frequency, high-consequence maneuvers within the AV Core team. You will drive problem definition, modelling, evaluation, integration, and deployment evidence across teams and partners.

Responsibilities include setting strategy, training policy models, and building data strategies for rare events, while ensuring robust verification and safe integration with the