Senior ML Infrastructure Engineer (Cortex Training)

Neura Market; Bellevue, Northern (WA, KY); Hybrid; USD 180,000 - 250,000; Full time

Job summary

Snowflake is seeking an Senior Software Engineer for Cortex Training to advance the ML platform that lets customers run demanding ML/AI workloads inside Snowflake. You will design and implement across the full stack, ensuring scalable, fault-tolerant GPU compute for post-training workflows, RL, and inference.

You will collaborate with Research and Platform teams to turn state-of-the-art techniques into reliable, enterprise-grade components, with a focus on performance, reliability, and cost

Qualifications

• 5+ years building and shipping production ML systems.
• Strong distributed systems and infrastructure foundation — Kubernetes in production.
• Familiarity with GPU and LLM infrastructure (PyTorch, DeepSpeed, Ray, CUDA/NCCL, vLLM).
• Ability to harden complex systems for reliability, throughput, and cost efficiency.
• BS in Computer Science or related field; MS/PhD a plus.

Responsibilities

• Design and build across the full stack — from public training APIs to the GPU data plane.
• Scale distributed systems to make GPU compute serverless with multi-tenant scheduling and fault tolerance.
• Drive end-to-end performance for training, inference, and RL loops under heavy load.
• Productionize research building blocks to deliver reliable, composable components at enterprise scale.

Skills

Production ML systems

Distributed systems

Kubernetes in production

GPU / LLM infra

Post-training / LLM modeling (bonus)

Education

BS in Computer Science or related field

MS/PhD a plus

Tools

PyTorch

DeepSpeed/FSDP

Ray

CUDA/NCCL

vLLM

Job description

Snowflake is seeking an Senior Software Engineer for Cortex Training to advance the ML platform that lets customers run demanding ML/AI workloads inside Snowflake. You will design and implement across the full stack, ensuring scalable, fault-tolerant GPU compute for post-training workflows, RL, and inference.

You will collaborate with Research and Platform teams to turn state-of-the-art techniques into reliable, enterprise-grade components, with a focus on performance, reliability, and cost