Enterprise AI Agent Platform
conceptA multi-agent orchestration layer for enterprise workflows — task planning, tool routing, and human-in-the-loop review built for reliability at scale.
- TypeScript
- Python
- LangGraph
- Postgres
A mix of production systems, open-source tools, and things I'm actively building toward.
A multi-agent orchestration layer for enterprise workflows — task planning, tool routing, and human-in-the-loop review built for reliability at scale.
A living reference for applied AI engineering — architecture patterns, evaluation frameworks, and postmortems from production systems.
Composable building blocks for AI product engineers — typed SDKs, evaluation harnesses, and prompt-versioning tools released under permissive licenses.
A reference implementation of retrieval-augmented generation built for correctness under load — chunking strategy, hybrid search, and grounded citations.
Tooling and servers for the Model Context Protocol — connecting AI systems to real data sources and actions safely.
Observability and evaluation infrastructure for teams running LLMs in production — tracing, regression testing, and cost monitoring.