Machine Learning Engineer, Synthetic Data at Wayve
United Kingdom
Posted
on-site · full_time · Visa sponsorship
Responsibilities:
- Post-train and iterate GAIA-class world models for rig transfer, pose transfer and related conditioning.
- Run and scale large-generation jobs (SDS/Flyte), manage GPU capacity, caches and generation correctness; debug visual artefacts and odometry issues.
- Convert generated outputs into training-ready artefacts: binarisation, metadata, sampling hooks and mix strategies for training stacks.
- Lead experiments to measure synthetic data impact on behaviour cloning, RL and evaluation suites; collaborate with driving-model owners on mix ratios and closed-loop evaluation.
- Work between research and platform engineering to ship checkpoints, generation pipelines and reproducible lineage for produced datasets.
Requirements:
- Applied ML experience with hands-on Python and PyTorch model training and evaluation.
- Background in generative models, neural rendering, world models or related video/3D synthesis techniques.
- Experience with GPU-scale inference, batching, memory/caching and production generation pipelines.
- Familiarity with data-centred training workflows, dataset provenance and metrics for synthetic-data quality.
- Strong debugging skills for visual/failure modes and ability to collaborate across research, simulation and platform teams.