Senior Software Engineer, Open Harness Engineering
Core
Building core libraries, evaluation systems, and reusable harness capabilities for open-source AI agent frameworks to ensure safety, speed, and reliability.
Role type
Senior Software Engineer (Open Source AI Agent Infrastructure)
Builds
Open-source agent harnesses, sandboxed execution environments, model-provider infrastructure, and agent benchmark systems.
Domain
AI Agents, Open Source Software, Developer Tools
Deliverable
production ML models
Required skills
Python, Rust, Go, C++, TypeScript, asynchronous execution, subprocess management, containerization, sandboxing, distributed compute, benchmark methodology, regression testing, open-source contribution, API development, trace analysis, observability.
Preferred skills
Experience with SWE-bench, Terminal-Bench, AgentBench, OpenTelemetry, OpenInference, structured event pipelines, debugging tools, published research in autonomous software engineering.
Technologies
Python, Rust, Go, C++, TypeScript, OpenTelemetry, OpenInference, CI/CD pipelines, Docker/Containers
Responsibilities
Define technical approaches for evaluating and improving open agent harnesses; build and operate scalable evaluation systems; diagnose agent failures from traces and logs; design controlled experiments and promotion criteria; ship harness and runtime improvements including upstream contributions; partner with research and infrastructure teams.
Seniority
Senior, hands-on IC with architecture design leadership
