Software Engineer, AI Runtime & Platform Services

What You'll Do

• Build and maintain the Python enterprise runtime around CrewAI: FastAPI services, Celery workers, Redis-backed state, execution APIs, and deployment-facing tools.
• Extend open-source CrewAI behavior for enterprise environments while preserving compatibility with upstream framework changes.
• Own production execution flows: crew and flow kickoff, status, retries, cancellation, checkpoint restore and fork, chat/session state, and human-in-the-loop resume paths.
• Build secure integration surfaces: JWT auth, signed webhooks, token refresh, file handling, secret fetching, and workload identity across AWS, GCP, and Azure.
• Improve observability across distributed execution: OpenTelemetry traces, structured logs, Sentry, event tracking, and debuggability across API, worker, and platform boundaries.
• Maintain strong test coverage for async/runtime behavior using pytest, mypy, ruff, mocks/fakes, and e2e deployment harnesses.
• Partner with the Agent Management Platform team on API contracts, versioning, enterprise client behavior, deployment status, and failure reporting.

Requirements

• Strong Python backend/platform engineering experience, especially building production services rather than only libraries.
• Experience with FastAPI or similar API frameworks, Celery or other job systems, Redis, Pydantic, and typed Python.
• Good instincts for distributed systems: retries, idempotency, async execution, status tracking, race conditions, and failure recovery.
• Comfort with auth and security-sensitive systems: JWTs, webhooks, signatures, secrets, IAM/workload identity, and least-privilege thinking.
• Practical observability experience: tracing, structured logging, metrics, Sentry/OpenTelemetry, and debugging multi-service failures.
• Ability to work at the boundary between an open-source framework and a hosted enterprise platform without creating brittle coupling.
• Strong testing habits and comfort with CI, package/version management, and release discipline.

Nice-to-Haves

• Experience operating AI/agent runtimes, workflow engines, or distributed task systems.
• Cloud platform experience with AWS ECS/ECR, Kubernetes, Helm, GCP/Azure identity, or secret managers.
• Experience with enterprise SaaS constraints: auditability, tenant isolation, customer environments, deployment rollbacks, and supportability.
• Familiarity with Rails/SaaS platforms is useful, but not required.