I build production AI systems and full-stack applications — deployed pipelines that handle errors, manage costs, and your team can maintain. Not prototypes.
Receipts: 1,000+ merged pull requests and 6,000+ GitHub contributions in the last twelve months, across three organizations.
What I've shipped recently:
• Release engineering for a multi-agent AI platform in classified defense environments: weekly release train, one signed and SBOM image digest from test to production, two human approval gates opened automatically as PRs, CVE remediation policy, Helm/ArgoCD/Terraform and Zarf/UDS packaging for air-gapped Kubernetes. 100+ merged PRs in that repo alone.
• Platform and agent-runtime code for the same product: FastAPI backend with per-gateway service identity, MCP-native runtimes for the Anthropic and Cohere SDKs, an outbound connector system (skills registry, MCP adapters, managed OAuth/API-key lifecycle, code and CLI commands), 70+ architecture decision records.
• A hub-and-spoke web platform generating production-ready sites across news, government data, community, and education..
• Cross-platform mobile with Expo/React Native: 58 merged PRs in the last year. 35+ native integrations, App Store published.
• hallow: open-source deterministic codebase intelligence for Python: dead code, duplication, complexity, architecture boundaries, SARIF output, CI audit mode.
Before freelancing: $14M cloud cost reduction at GE Digital. Co-founded the OpenStack Foundation (valued at $5B). Unified 40+ team members across 5 countries at a global infrastructure company. 25+ years across enterprise (GE Digital, Yahoo, Walmart, VMware), government, defense, and startups.
Tech: Python, TypeScript, FastAPI, Next.js, React, React Native/Expo, Claude API, Vercel AI SDK, MCP, LLaMA, Portkey, Langfuse, PostgreSQL, Redis, Docker, Kubernetes, Helm, ArgoCD, Terraform, AWS, Azure, GitHub Actions, GitLab CI, Vercel.
I send pull requests, not slide decks.
Sean R. earns an estimated $9.9k/mo. That's 8.6× the typical freelancer and more than 99.88% of everyone we track.