ML Platform Engineer — Distributed Training & Infra

Breakout Ventures; United States; On-site; USD 180,000 - 260,000; Full time

Job summary

Noetik is seeking highly technical ML engineers to join a hybrid South San Francisco team building scalable AI systems for oncology. You will design, train, and deploy production-grade workflows powering large foundation models and data-driven drug discovery.

Join a cross-disciplinary group focused on reproducibility, rapid iteration, and close collaboration with ML researchers. The role emphasizes scalable infrastructure, observability, and high-performance GPU workloads in a hybrid setup.

Qualifications

• Strong software engineering fundamentals for production ML systems.
• Experience with distributed model training in large-scale environments.
• Reading, profiling, debugging, and modifying model training and inference code.

Responsibilities

• Design, build, and evolve systems for distributed model training and batch inference.
• Develop workflows for experiment tracking, model evaluation, and registry management.
• Profile and optimize training and inference performance across PyTorch, data loading, GPU utilization, and memory.
• Build researcher-facing CLIs, SDKs, and job submission tools for scalable runs.
• Develop and operate cost-efficient GPU infrastructure using Kubernetes, AWS, Ray, and Terraform.
• Collaborate with ML researchers to translate needs into platform improvements.
• Use AI development tools to accelerate implementation, debugging, profiling and testing.

Skills

Distributed training

PyTorch

GPU systems

Kubernetes

Cloud infrastructure

Profiling tools

Debugging complex systems

Collaboration with researchers

Tools

Terraform

Docker

Ray

MLflow

Job description

Noetik is seeking highly technical ML engineers to join a hybrid South San Francisco team building scalable AI systems for oncology. You will design, train, and deploy production-grade workflows powering large foundation models and data-driven drug discovery.

Join a cross-disciplinary group focused on reproducibility, rapid iteration, and close collaboration with ML researchers. The role emphasizes scalable infrastructure, observability, and high-performance GPU workloads in a hybrid setup.