Software Engineer - Compute Infrastructure for Frontier AI

CV in; Northern (KY); Hybrid; USD 180,000 - 240,000; Full time

Job summary

OpenAI is seeking a Software Engineer for Compute Infrastructure to build and optimize the platform that powers frontier AI. You will design, provision, and operate large-scale systems connecting GPUs, CPUs, networks, storage, and orchestration to run demanding workloads.

Ideal candidates have distributed systems experience, HPC familiarity, and a track record of reliability engineering, observability, and performance tuning at scale.

Qualifications

• Strong software engineering skills and experience building, operating, or improving production infrastructure systems are essential.
• Experience in distributed systems, operating systems, networking protocols, RDMA, NCCL or collective communication, storage, Kubernetes, scheduling, observability, reliability engineering, high‑performance computing, GPU infrastructure, CaaS, agent infrastructure, hardware‑aware performance optimization, benchmarking, developer experience, or infrastructure tooling is required.

Responsibilities

• You will build and deeply optimize reliable system software for large-scale compute systems that run AI workloads.
• You will design and operate infrastructure across accelerators, CPUs, NICs, switches, networking protocols, storage, data centers, cluster orchestration, scheduling, and fleet health.
• Profiling, benchmarking, and optimization of training workloads across compute, memory, storage, networking, NCCL and collective communication will form part of your daily work.

Skills

Distributed systems

High-performance computing

Observability

Reliability engineering

Tools

Kubernetes

NCCL/GPUs

RDMA

Observability tooling

Job description

OpenAI is seeking a Software Engineer for Compute Infrastructure to build and optimize the platform that powers frontier AI. You will design, provision, and operate large-scale systems connecting GPUs, CPUs, networks, storage, and orchestration to run demanding workloads.

Ideal candidates have distributed systems experience, HPC familiarity, and a track record of reliability engineering, observability, and performance tuning at scale.