🔎
Current Focus: RAG & Context Engineering
Active practical development
My recent development work has focused on a practical question: how can a business make its existing data useful to AI without replacing the systems that already run the company?
I am working through RAG and context engineering using ASP.NET Core, Microsoft Foundry (formerly Azure AI Foundry) and Azure AI Search. Context engineering decides which instructions, rules, examples, tool results, conversation state and retrieved evidence a model needs. The RAG pipeline prepares authorised SQL data, documents and APIs through extraction, chunking, metadata and embeddings, then uses vector, semantic and hybrid retrieval to supply grounded evidence with citations.
The existing application and database remain the source of truth. The retrieval layer must preserve meaning, prefer current authoritative information, enforce permissions and recognise when reliable evidence cannot be found. A trustworthy AI answer begins with trustworthy context. This is active hands-on study, prototyping and technical writing built on my established enterprise data, API, security and workflow experience; I do not present it as long-term production AI experience.
Context Engineering Microsoft Foundry Azure AI Search Retrieval-Augmented Generation Embeddings Vector, Semantic & Hybrid Search Chunking & Metadata SQL & Document Ingestion Retrieval & Reranking Grounded Answers & Citations Security & Access Filtering Advanced RAG Patterns
Related work
Complete AI Engineering learning journey → Context engineering guide → Reliable LLM application guide → Embeddings and semantic search guide → Complete RAG ingestion pipeline guide → RAG chunking strategies guide → RAG retrieval strategies guide → Advanced RAG patterns guide → Business Data to RAG guide → RAG and vector databases guide → Azure AI Foundry guide →