SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Wayve is building the leading AI platform for autonomous driving using end-to-end AI that enables vehicles to learn from real-world experience. The Performance Architecture team is part of Wayve's AI Performance org and focuses on making Wayve's AI workloads faster and more efficient across training and cloud inference.
In this role, you will identify, quantify, and deliver optimizations across training and cloud inference workloads. Your day-to-day responsibilities include profiling workloads to find bottlenecks, building optimizations that work across multiple targets rather than one-off fixes, and tracking gains with clear benchmarks. You'll work closely with Research and model teams to integrate performance engineering into their development cycle, and partner with platform teams on cloud GPU hardware strategy.
Key responsibilities:
- Profile ML workloads across training and cloud inference using system and kernel-level profilers to identify bottlenecks
- Design and implement efficiency improvements to maximize MFU, throughput, and utilization (e.g., parallelism, compilation, mixed precision, caching)
- Build reusable, cross-target optimizations (kernels, data loaders, frameworks like Triton) rather than one-off, per-workflow fixes
- Design and implement benchmarking tools to track efficiency gains and catch regressions on priority workloads
- Inform cloud GPU hardware strategy and readiness in partnership with platform teams
- Build a culture of performance optimization with Research and model teams
Wayve operates a hybrid working model combining in-person collaboration in dedicated office spaces with focused remote work time. Main hubs include London, Sunnyvale, Yokohama, Herzliya, Vancouver, and Leonberg.
Requirements:
- 10+ years of industry experience driving performance engineering across ML systems, GPU compute infrastructure, distributed platforms, or similar
- Experience optimizing large-scale workloads on GPU compute clusters for training, inference, or both
- Demonstrated ability to write, report, and track performance benchmarks in an open and accessible way
- High-quality, well-structured, and tested Python code
- BS or MS in Machine Learning, Computer Science, Engineering, or related technical discipline, or equivalent experience
Nice to have:
- Experience with concurrent, parallel, and distributed computing
- Experience optimizing inference serving systems (latency, throughput, batching, caching)
- Experience using NVIDIA Nsight Systems or other system profilers
- Experience implementing GPU kernels (CUDA, Triton, etc.)
- Knowledge of computing fundamentals—what makes code fast, secure, and reliable
