Software Engineer - Cortex Training

• Design and build across the full stack - from the public training APIs and SDK through the control plane to the GPU data plane.
• Scale the distributed systems that make GPU compute serverless - multi-tenant scheduling, placement, and capacity-aware routing across regional GPU pools, with fault tolerance built in.
• Drive end-to-end performance at scale - keep the training, inference, and RL loops fast and the data plane responsive under heavy concurrent load, with GPUs kept saturated.
• Productionize research building blocks - partner with Snowflake Research to turn state-of-the-art training and inference techniques into reliable, composable components customers can run at enterprise scale.

• 3 + years (Intermediate)

  6+ years (Senior) building and shipping production ML systems

• Strong distributed systems and infrastructure foundation - designing scalable, fault-tolerant services and operating them on Kubernetes in production.
• Familiarity with GPU and LLM infrastructure - e.g., PyTorch, DeepSpeed/FSDP, Ray, CUDA/NCCL, vLLM; able to debug across the data, infrastructure, and GPU layers.
• Demonstrated ability to harden complex systems for reliability, throughput, and cost efficiency.
• BS in Computer Science or a related field (MS/PhD a plus).
• (Bonus) Hands-on LLM post-training / modeling experience - the strongest candidates pair deep infra skills with real post-training intuition.

Snowflake is growing fast, and we’re scaling our team to help enable and accelerate our growth. We are looking for people who share our values, challenge ordinary thinking, and push the pace of innovation while building a future for themselves and Snowflake.

The Snowflake ML Platform team’s mission is to let customers run their most demanding ML/AI workloads inside Snowflake. Cortex Training is our LLM post-training platform: it turns scarce, expensive GPU capacity into a simple, composable service, so customers can adapt open-weight foundation models to their own business problems while we handle the hard distributed-systems parts, including scheduling, orchestration, multi-node training and inference, fault tolerance, and throughput.

The platform already runs post-training at scale. Under the hood, it decouples GPU computation from the training loop and exposes it as primitive APIs that compose into everything from SFT to full RL workflows. You’ll work alongside a team that ships fast & sweats reliability and the researchers behind DeepSpeed. We’re looking for an engineer who thrives in the ML infrastructure layer and brings a solid understanding of LLMs and post-training to help us scale and grow it.