Member of Technical Staff - Inference

Sail Research•San Francisco, CA

5d•Onsite

About The Position

Sail builds the world's most efficient software for inference (processing LLM tokens) and agent hosting (cloud VMs). Together, our technologies allow our customers to deploy AI agents at large scale to do the most challenging work. In this role, you'll own token processing down to the lowest layers of the stack. You'll do things like: develop a new request scheduling strategy, achieve better communication/computation overlap, investigate novel schemes for increasing cache hit rates, or identify a better way to benchmark inference performance.

Requirements

Strong understanding of core LLM mechanics, like KV cache, mixture-of-experts, prefill vs. decode phases.
Interest in MLSys research - great ideas like speculative decoding and sparse attention come from research, that we need to follow closely.
Familiarity with modern, tile-based GPU programming, e.g. Triton, CUTLASS, ThunderKittens, etc. Or an interest in learning these!
Great interpersonal and technical communication.
Please don't use LLMs to write prose. We desk-reject slopful cover letters and resumes.

Responsibilities

Modify and extend state-of-the-art inference engines like vLLM and SGLang, and work on our own internal engine.
Understand every microsecond of GPU time spent during a forward pass. You'll be able to explain every kernel launch on an nsys profile.
Design and implement exotic parallelism schemes to work with "interesting" hardware topologies.
Write and debug GPU kernels to excel in specific regimes, such as cascade attention.