Key Notes
Gaia, Wayve’s video world model, predicts future frames from driving footage and generates synthetic scenarios, including rare or safety-critical events. This Staff ML role owns large-scale training and improvement of in-house foundation models. Large-model training, architecture and modern ML engineering experience are essential; relevant industry experience is typically four–five-plus years. The full-time London role follows Wayve’s hybrid policy.
What You'll Work On
- Lead large-scale video or adjacent foundation-model training runs from experimental design through reliable, production-grade execution across Wayve’s infrastructure.
- Shape Gaia’s model architecture and training strategy through first-principles analysis instead of relying on off-the-shelf model application.
- Improve Gaia’s world-model capabilities for synthetic scenario generation and downstream driving-model evaluation or training across supported development workflows.
- Partner with research, applications, simulation engineering and cloud-infrastructure teams while setting technical standards through mentorship, review and direction.
Why This Role Matters
- Gaia’s synthetic scenarios include rare or safety-critical driving events; improvements to the video world model expand the simulation capability available for downstream driving-model evaluation and training.
- Gaia experiments must run at frontier-model scale; coordinated model and infrastructure work connects experimental design with production-grade training execution and reliable end-to-end development.
What They Are Looking For
- Experience: In-depth ownership of large-scale language, video or other foundation-model training is essential.
- Technical: Strong model-architecture knowledge and hands-on engineering with modern ML stacks such as PyTorch are essential.
- Experience: Relevant applied industry experience, typically 4–5+ years, is essential; advanced degrees are valued but not required.
- Preferred: World models, video generation, long-horizon prediction or distributed training experience is desirable.
- Preferred: Technical leadership, mentoring and direction-setting across an ML area are desirable.