Odyssey
Member of Technical Staff, ML Performance
Palo Alto +2 · Hybrid · Staff · Full-time · Posted Mar 11, 2026
- Location
- Palo Alto
Zurich
London - Experience
- 8+ years of software engineering experience, with significant work in ML performance.
About the job
Optimize machine-learning models and infrastructure for fast, reliable, scalable training and real-time inference.
What you'll do
- Optimize models for real-time use at large scale and develop distributed training strategies for large GPU clusters.
- Identify performance bottlenecks and stability issues in training and serving environments.
- Design tools, frameworks, and systems to improve performance across model development and inference infrastructure.
What they ask for
- Proficiency with PyTorch (or TF/JAX), Triton, and NVIDIA GPU ecosystems and optimization stacks.
- Deep knowledge of modern ML architectures and performance optimization, particularly distributed training and inference.
- Track record of owning projects end to end.
- Metric-based approach and ability to learn new skills as needed.
Tools
PyTorch, TF, JAX, Triton, NVIDIA GPU
AI focus
Improve performance, efficiency, reliability, and scalability of ML model training and inference.
Model-written summary · the original posting is authoritative