Senior Software Engineer, Cortex LLM Training Platform

Snowflake; Bellevue (WA); On-site; USD 200,000 - 288,000; Full time

Benefits offered by this job

Medical, dental, vision

401(k) retirement plan

Paid holidays and time off

Job summary

Snowflake is seeking an experienced Senior Software Engineer for Cortex Training to scale their ML platform. You will design the full stack from APIs to GPU data plane, ensuring fast, reliable training and inference at enterprise scale.

You will work with DeepSpeed, PyTorch, and GPU infrastructure, partnering with researchers to turn state-of-the-art methods into production-ready components. A strong distributed-systems background is essential.

Qualifications

• BS in Computer Science or related field; MS/PhD a plus.
• 3+ years building production ML systems; 6+ years for senior level.
• Strong background in distributed systems and GPU infrastructure.

Responsibilities

• Design and build across the full ML stack from public APIs to GPU data plane.
• Scale distributed GPU compute with multi-tenant scheduling and fault tolerance.
• Optimize training, inference and RL loops for high concurrency and throughput.
• Productionize research building blocks with reliable, composable components.

Skills

Distributed systems

Kubernetes

PyTorch

DeepSpeed

Ray

Education

BS in Computer Science

Tools

CUDA/NCCL

vLLM

DeepSpeed/FSDP

Job description

Snowflake is seeking an experienced Senior Software Engineer for Cortex Training to scale their ML platform. You will design the full stack from APIs to GPU data plane, ensuring fast, reliable training and inference at enterprise scale.

You will work with DeepSpeed, PyTorch, and GPU infrastructure, partnering with researchers to turn state-of-the-art methods into production-ready components. A strong distributed-systems background is essential.