Job Details

Machine Learning Engineer

Tech start-up

San Fransisco based

We've partnered with one of the most ambitious and technically rigorous AI research labs in the world. Based in San Francisco, this team is building foundation models entirely from scratch.

They are now hiring ML Infrastructure Engineers to design and scale the systems that power large-scale, distributed model training. If you've built infrastructure that runs across hundreds of GPUs, thrive under technical complexity, and want to work side-by-side with elite AI researchers — this is the role.

Key Responsibilities:

• Build and scale distributed training systems for large-scale model training across LLMs, vision, and robotics.
• Set up and run large-scale training across many GPUs using tools like Kubernetes, DeepSpeed, and FSDP.
• Troubleshoot system issues (GPU errors, network problems) and build tools to monitor and recover from failures.
• Optimize PyTorch pipelines, sharding, and sampling strategies.
• Collaborate closely with researchers to support novel model training at scale.

Requirements:

• 3–15 years in ML infrastructure, systems, or research engineering roles.
• Proven experience scaling distributed training for large models.
• Strong with PyTorch, CUDA, NCCL, Kubernetes.
• Familiar with setting up distributed training clusters.
• Deep understanding of PyTorch dataloaders, data sharding, and sampling.
• Strong communicator with a collaborative, mission-driven mindset.

This is a fully in-person role based in San Francisco, it's ideal for engineers excited to build at the edge of what's possible in AI.