The point where experts and best companies meet
Tesla Staff Performance Optimization Engineer AI Infrastructure United States, California, Palo Alto 871180934
06.04.2025
Share
What You’ll Do What You’ll Bring
- Work with a wide variety of teams at Tesla to accelerate time-to-market for new ML models
- Design, implement, and deploy low-overhead instrumentation methods for troubleshooting performance issues
- Analyze collected telemetry, identifying bottlenecks and designing practical solutions to overcome those bottlenecks
- Develop data-driven performance improvements to existing software pipelines
- Working with each team to validate these performance improvements and incorporate them into production training runs
- Help make Tesla's ambitious AI-related products and services a reality
- A deep understanding of the internals of GPU-based training and inferencing workloads, especially handoffs of data and computation between host CPUs and GPUs
- Real-world knowledge of supporting languages and libraries used in large-scale AI training runs (CUDA/ZLUDA, OpenCL, PyTorch, Tensorflow, GPUDirect and other RDMA-enabling services)
- Experience developing and tuning low-level software using languages like C, x86 assembly, and Rust
- Practical experiences using different performance analysis techniques (profiling, tracing, simulative analysis) and when each should be applied
- Excellent spoken and written communication skills, including the ability to concisely communicate data-driven root causes of performance issues and how they can be remedied
- An irrational love for high-performance computing and extracting the maximum number of productive FLOPs from modern AI-oriented architectures
These jobs might be a good fit
