All services
AI engineering
I take LLM features from prototype to production. Multi-agent systems with LangGraph, hybrid RAG with source grounding, plus evaluation and observability. The entry point is the LLM readiness check.
What's in the work
Multi-agent systems with LangGraph
Hybrid RAG with source grounding
Evaluation and observability (Langfuse, RAGAS)
MCP servers when the model needs tools
Hook into the existing app without rebuilding it
Technologies
PythonLangGraphLangChainOpenAI / Azure OpenAIpgvectorMilvusFastAPIMCPRAGLangfuse
Related work
AI Audit Assistant for Certification & Compliance
An assistant for ISO audits: LangGraph agents and RAG over the standards. Raw notes become findings you can tie to a clause.
PythonLangGraphFastAPIAzure OpenAIpgvectorNext.js
More details
AI Workspace for Food Scientists
A workspace for food scientists, with separate agents for recipes, ingredients, suppliers and rules. Not one chatbot.
PythonLangGraphLangChainFastAPIKnowledge Graphs
More details
AI Chatbot for Knowledge Platform
A chatbot with a RAG pipeline over 500+ video and PDF items. An MCP server is attached.
PythonLangChainOpenAIMilvusFastAPIMCP
More details