I build production-grade AI agent systems in .NET: multi-agent orchestration, MCP/A2A integrations, retrieval over large codebases, and LLM grounding. There are 25+ years of professional software engineering behind that work.
Website | Download CV | Deutsch
- Agentic AI systems in .NET/C#
- Multi-agent orchestration, guardrails, workflow DSLs
- MCP/A2A integrations and tool-driven LLM workflows
- Retrieval over codebases, RAG, and LLM grounding
- Enterprise backend architecture, APIs, Azure, and production delivery
- nexus.net - multi-agent orchestration engine for .NET with graphs, guardrails, memory, MCP/A2A/AG-UI, and workflow DSL.
- code-explorer - token-efficient codebase retrieval for AI agents with tree-sitter AST indexing, BM25 search, and fast symbol lookup.
- linxmd - package manager for AI workflow artifacts such as agents, skills, and workflows.
- web-explorer - LLM-ready web search and extraction toolkit for reliable grounding.
- ladybug.net - idiomatic .NET bindings for the Kuzu embedded graph database.
- beads.net - local-first issue tracker in a single SQLite file, available as .NET library and CLI.
AI / Agents: MCP, A2A, RAG/Retrieval, LLM grounding, ONNX, agent workflows
Backend: C#, .NET, ASP.NET Core, APIs, SQL Server, PostgreSQL, SQLite
Delivery: Azure, Docker, GitHub Actions, CI/CD, observability-minded architecture
Product reach: TypeScript, Node.js, React, Flutter, Electron
Remote work in German or English:
- fixed-scope development projects
- team advisory and architecture reviews
- retainers for AI/.NET product development
- production hardening of AI prototypes
I work best with clear goals, async-friendly communication, and ownership from architecture to delivery.
- Website: knaack.it
- LinkedIn: linkedin.com/in/denisknaack
- Stack Overflow: stackoverflow.com/users/479659




