jobs

AI Systems Engineer, Codex Agents

The Codex Core Agents team builds the agent harness that turns model capability into real-world action. The team owns the systems surrounding the model, including prompting, model-output interpretation, safe tool execution, sandboxed enviro

Company

OpenAI

Location

San Francisco, United States

Status

Closed

Posted

2026-09-13T13:31:38.707+00:00

OpenAI's Codex Core Agents team is hiring an AI Systems Engineer in San Francisco, hybrid. The role builds the harness around the model: the execution loop that lets agents interpret model outputs, call tools and run code safely, plus sandboxing and isolation, evaluation and debugging systems, prompt and tool-use experiments, and observability across backend, inference and GPU fleets. The harness is meant to be trainable, measurable and shared with open-source users. The posting asks for distributed systems, infrastructure, sandboxing, virtualization or ML systems experience; LLM applications, evaluations, profiling and Rust are valuable.