Fiona’s Experience Agent (CV + RAG demo)

Purpose

A retrieval-augmented agent over Fiona’s detailed CV and portfolio. It rewrites your question into a focused search query, pulls the most relevant roles and projects, and returns concise, skimmable answers highlighting how her experience maps to your needs. The agent never invents facts and will say when it can’t find anything relevant in the source document.

Capabilities

• Answer questions about Fiona’s roles, projects, and responsibilities
• Summarise experience for specific roles or industries
• Highlight transferable skills and relevant achievements
• Map user requirements to relevant experience
• Handle questions phrased as if speaking directly to Fiona (e.g. 'Where have you worked?')

Techniques

• Retrieval-augmented generation (RAG) over a CV/experience document
• Semantic search with vector embeddings
• Query rewriting to improve recall over the document
• Guardrailed generation (no fabrication outside the source)

Technology

• Node.js + Express backend
• Next.js frontend
• OpenAI GPT-4.1-mini for reasoning and generation
• OpenAI text-embedding-3-small for semantic search
• Custom per-agent daily token budgeting and override mechanism

Guardrails

• Answers are grounded only in the CV/experience document
• Does not invent roles, dates, companies, or achievements
• Clearly states when it cannot find relevant information
• Treats any 'you' phrasing as referring to Fiona, not the user
• Does not process or store end-user personal data beyond questions themselves
• All usage (questions, token consumption, password overrides) is logged for monitoring