Software Engineer (Infrastructure)

Exa; San Francisco (CA); On-site; USD 120,000 - 160,000; Full time

Job summary

A cutting-edge tech company in San Francisco seeks infrastructure engineers to enhance the tooling and systems that power its AI applications. Responsibilities include building GPU orchestration, scaling cloud batchjob systems, and designing efficient scheduling software. Candidates should have experience with large-scale infrastructure and a strong focus on reliability and observability. This position is in-person, and international sponsorship is available.

Qualifications

• Experience with large-scale infrastructure, such as GPU clusters or cloud batchjob systems.
• Obsessive focus on reliability, observability, and optimization.

Responsibilities

• Build GPU cluster orchestration.
• Scale AWS batchjob systems to handle extensive jobs.
• Design GPU scheduling software for maximum utilization.
• Develop observability tooling for production systems.

Skills

Designing and operating large-scale infrastructure

Reliability and observability mindset

Optimization across the entire stack

Tools

Kubernetes

AWS batchjob systems

GPU scheduling software

Job description

Exa is building a search engine from scratch to serve every AI application. We build massive‑scale infrastructure to crawl the web, train state‑of‑the‑art embedding models to index it, and develop super high performant vector databases in Rust to search over it. We also own a $5M H200 GPU cluster that regularly lights up tens of thousands of machines.

Our Infrastructure Team builds the underlying tooling and infrastructure that powers all Exa's systems. Basically, we need more infra engineers to build the machine that builds the machine so that we can move as fast as possible as an engineering org. That could mean building GPU cluster orchestration in Kubernetes, map‑reduce batchjobs on Ray, or the best observability tooling in the world.

Desired Experience

• You have experience designing and operating large-scale infrastructure - GPU clusters or large Kubernetes clusters or cloud batchjob systems
• You bring an obsessive mindset — always thinking about reliability, observability, and optimization across the entire stack.

Example Projects

• Build the Kubernetes orchestration on a $20m GPU cluster
• Scale our AWS batchjob system to handle map reduce jobs over 10s of thousands of machines
• Design GPU scheduling software so we max out our cluster utilization
• Build observability into our production systems

This is an in-person opportunity in San Francisco. We are open to sponsoring international candidates (e.g., STEM OPT, OPT, H1B, O1, E3).