Senior Software Engineer Model Hosting- remote

Senior Software Engineer - Model Hosting

The Model Hosting team sits at the heart of our infrastructure, working deeply within our bare metal and Kubernetes stack to ensure every bit of compute is used to its fullest. If you're excited about squeezing maximum performance out of hardware and building the systems that serve large language models at scale, we'd love to talk.

About the Role
You'll help design and operate the platform that serves our models in production. Our stack spans several languages, each chosen for the job it does best: Go powers our gateway layer, Rust handles fast real-time decision-making, and Python supports model dependencies that require it. Kubernetes orchestrates all model deployments across the company.

Your work will touch some of the most interesting problems in high-performance model serving, including:
RDMA technologies (InfiniBand and RoCE) across multiple nodes
Disaggregated serving and KV cache offloading
Quantization, speculative-decoding, MTP, and other inference performance optimizations
Working with modern inference engines such as SGLang, vLLM, and others

What We're Looking For

We are primarily looking for someone with 1-3 years of experience hosting models (a fairly new field), but who also has 5-10+ years of broader software engineering experience.

Experience serving LLMs in production
Strong programming skills in Python and/or Go
A track record of designing and operating highly scalable, highly available distributed services
Working knowledge of InfiniBand, RoCE, or other high-performance networking

Nice to Have
Contributions to vLLM, SGLang, TRT-LLM, or other NVIDIA ecosystem open-source projects
Deep understanding of Kubernetes
Experience developing in Rust
Familiarity with storage over RDMA