ML Infrastructure Engineer (LLM)

Location: Bellevue, Washington, United States

Salary: $200,000–$287,500 per year

Company

Snowflake develops a cloud data platform, including ML infrastructure for running demanding AI workloads at enterprise scale.

What you will do

• Design and build the full stack of Cortex Training, from public training APIs and SDKs to the control plane and GPU data plane.
• Scale multi-tenant scheduling, placement, and capacity-aware routing across regional GPU pools.
• Build fault-tolerant distributed systems for serverless GPU computing.
• Optimize training, inference, and reinforcement-learning loops under heavy concurrent load while keeping GPUs saturated.
• Partner with Snowflake Research to productionize state-of-the-art training and inference techniques.

Requirements

• 3+ years of experience for the intermediate level or 6+ years for the senior level building and shipping production ML systems.
• Strong distributed-systems and infrastructure foundation, including scalable, fault-tolerant services operated on Kubernetes in production.
• Familiarity with GPU and LLM infrastructure such as PyTorch, DeepSpeed or FSDP, Ray, CUDA or NCCL, and vLLM.
• Ability to debug across data, infrastructure, and GPU layers and harden systems for reliability, throughput, and cost efficiency.
• Bachelor's degree in Computer Science or a related field.

Nice to have

• Master's degree or PhD.
• Hands-on LLM post-training or modeling experience.

Culture & Benefits

• Work in an AI-focused engineering environment with an experimental approach to solving problems.
• Collaborate with researchers and engineers on production ML infrastructure.
• Build services for demanding enterprise-scale workloads.