Staff Machine Learning Engineer, Vision Models
Core
Build computer vision and scene understanding models to measure the performance of Wayve Driver offline, adapting on-vehicle architectures and foundation models for offline use to assess driving behavior and define ground truth.
Role type
Staff Machine Learning Engineer (Vision Models)
Builds
Offline scene understanding models for autonomous driving validation
Domain
Autonomous driving / Computer vision
Deliverable
production ML models
Required skills
Deep learning model training and shipping, Transformer-based and multimodal/VLM architectures, Foundation model adaptation and fine-tuning, Large-scale distributed training, Python and PyTorch, Cross-functional technical leadership, Model benchmarking and error analysis
Preferred skills
3D scene understanding and representation learning, Offboard/offline modelling techniques, Autonomous vehicle deployment and closed-loop validation, Fleet-scale data infrastructure
Technologies
PyTorch, Transformers, VLMs, Foundation Models
Responsibilities
Build, train, and fine-tune scene understanding models for offline measurement; Improve model performance and diagnose failure modes across platforms and conditions; Leverage offline advantages like higher compute and bidirectional temporal context; Benchmark models and define quality bars to steer iteration; Ensure benchmarked results are statistically defensible for safety cases; Mentor team members and align priorities across sites
Seniority
Staff, hands-on IC with technical leadership
