As an Evaluation Engineer, you’ll add new game features to existing game environments, and make brand new environments available. You’ll work collaboratively to develop and integrate environments and carry out performance optimisation to ensure that our games run fast in our data centres. We take time to design and implement our software carefully as it’s likely we’ll live with it for many years to come so we welcome new ideas and apply thoughtfulness to everything we do. We take code reviews, unit and integration testing seriously to ensure high code quality and robustness.
As part of the Research Group, the Evaluation Team is tasked with measuring the performance of AI agents created by researchers. This is done by providing challenges and tasks for agents to solve. These range from classic arcade games to first-person games using modern 3D engines to bespoke mini-games tailored to answer specific research questions. The output from the Evaluation Team is used by the entire organisation to track progress and is crucial in helping steer DeepMind’s research forward.
- Strong software engineer with extensive experience in software design
- 3+ years C++ programming experience; professional or otherwise
- Python and/or Lua experience ideal, but not essential
- Experience with gameplay, engine, graphics or physics programming
- Experience with profiling and performance optimisation
- Basic understanding of maths and statistics
- Ideally experience with concurrent or distributed software architecture
Competitive salary applies.