Machine Learning Engineer (Synthetic Data)

Overview

As a Machine Learning Engineer at Wayve, you will advance end-to-end autonomous driving research by building and scaling world-model architectures and the synthetic data they produce. You will bridge ML research with engineering to enable high-throughput generation and efficient training pipelines. Your work directly accelerates AV development by integrating synthetic experience into real-driving training. This role offers ownership of a core capability within a high-trust, collaborative team shaping the future of autonomous vehicles.

Pay / Benefits

• hybrid working policy (office + home)
• office in London
• core working hours
• flexible schedule
• collaborative culture
• opportunity to shape autonomous driving tech

Responsibilities

• Post-train and iterate GAIA-class world models for synthetic-data capabilities (rig transfer, pose transfer, conditioning).
• Own the generation loop: configure large-scale GPU inference and prepare training-ready artefacts with clear lineage.
• Land synthetic data in driving-model training (behavior cloning, reward models, RL) and evaluate on-road impact.
• Diagnose and fix geometry, calibration, and controllability failures affecting training-grade video.
• Improve throughput and yield through inference optimisations and self-serve workflows.
• Expand coverage to new vehicle platforms and safety-critical scenarios.
• Collaborate with world-model researchers, infra, and driving-model owners to maintain a single generation-evaluation-training system.

Key requirements

• 4+ years in applied ML / research engineering with training and shipping neural nets.
• Strong Python and PyTorch; GPU training, debugging, and reading model code.
• Hands-on experience with video, generative, or world models (diffusion / flow-matching / autoregressive video, neural rendering).
• Working knowledge of cameras and 3D geometry (intrinsics/extrinsics, warps, reprojection).
• Evidence of using generated or simulated data in downstream models with impact measurement.
• Ability to run generation/training at multi-GPU scale and ensure reliability.
• Collaborative, experimental mindset; capable of owning a capability end-to-end.

• collaborative
• experimental
• ownership
• Python
• PyTorch
• GPU training

…