vigads ai

This site isn’t a static resume. You can ask about my experience, skills, and projects — answers come from a RAG pipeline over my own knowledge base.

I wanted one project that mixes frontend craft with AI engineering. Building my own assistant was the best way to learn RAG for real.

Website
  • Node.js
  • Fastify
  • Ollama
  • LangChain
  • Neo4j
  • RAG
  • Qwen 3

Contributions

  • Designed the RAG architecture end to end
  • Built the Fastify chat API with streaming
  • Set up Neo4j vector search and embeddings
  • Integrated LangChain and local Ollama models
  • Created the Markdown knowledge ingestion pipeline

Features

  • Chat assistant about my career and work
  • Semantic search over a personal knowledge base
  • Local LLM inference with Ollama
  • Streaming responses in the UI
  • Semantic cache for repeated questions

Fundamentals

  • Retrieve before generate — ground answers in real context
  • Keep models local when privacy and cost matter
  • Treat the knowledge base as a product surface, not a dump