Member of Technical Staff, ML Performance

The role

This person will optimize models for real-time use and design distributed training strategies on large GPU clusters. They will work to ensure model speed, reliability, and scalability in training and inference phases.

What you will do

• Optimize models that will be used in real-time by hundreds of thousands of users
• Design and implement distributed training strategies to reduce training time and resource consumption on large GPU clusters
• Partner with ML researchers and engineers to ensure model architectures are highly performant
• Develop tools to identify performance bottlenecks and stability issues in training and serving environments
• Pioneer innovative approaches, frameworks, and system designs to enhance performance metrics
• Have significant autonomy in technical decisions
• Use the latest-generation GPUs

Required qualifications

• 8+ years of software engineering experience, with significant work in ML performance
• Deep insight into modern machine learning architectures
• Natural instinct for performance optimization, particularly distributed training and inference
• Track record of owning projects end to end
• Problem-solving mindset with the ability to acquire new skills as needed
• Proficiency with PyTorch (or TF/JAX) and Triton
• Proficiency with NVIDIA GPU ecosystems and optimization stacks
• Highly metric-based

About the company

Odyssey is an AI lab pioneering general world models. The founders previously pioneered self-driving cars and have brought together a research team from DeepMind, Tesla, Waymo, Meta, Apple, and Wayve.