Software Engineer, Model Inference, DeepMind
Summary
Build and optimize ML serving infrastructure and inference backends to accelerate model releases and improve performance for DeepMind’s AI workloads.
- Automate tasks and reduce redundancies
- Build performant tests for model releases
- Collaborate with research teams on modeling approaches
- Deliver serving infrastructure for ML
- Implement caching mechanisms
- Improve model release velocity
- Optimize inference backends for accelerators
- Optimize serving infrastructure performance and scale
- Profile hardware and software for performance bottlenecks
- Understand serving frameworks and preprocessing pipelines
- Use roofline analysis to optimize ML workloads