Key Notes
Foundation models give Waymo’s Simulator realistic agents, roads and weather for Driver testing and training. This London hybrid role designs experiments and realism metrics, trains models and integrates research into production with Oxford and Mountain View teams. Five or more years in applied deep learning and coding/design are preferred. Expected base salary is £123,000–£129,000, with bonus and equity subject to eligibility.
What You'll Work On
- Train large foundation models that represent vehicles, pedestrians, cyclists, roads, traffic controls and weather in Waymo simulation.
- Design experiments and quantitative metrics that measure the realism and steerability of simulated autonomous-driving worlds.
- Integrate trained foundation models into Waymo's core simulator and its downstream Driver evaluation and training applications.
- Productionise research ideas with London, Oxford and Mountain View teams and participate in hiring research engineers.
Why This Role Matters
- Waymo Driver testing and training draw on simulated agents, roads and weather; improved realism and controllability provide a more representative virtual environment for evaluating driving behaviour.
- Waymo’s evaluation and training products can use foundation-model research once the models are integrated into the simulator and its downstream applications.
What They Are Looking For
- Experience: At least five years of applied deep-learning experience is preferred.
- Technical: At least five years of software coding and design experience is preferred.
- Experience: Solving complex production problems with current machine-learning techniques is preferred.
- Experience: Taking research ideas into production systems is preferred.
- Technical: Data-analysis or data-science expertise is preferred.