AI Agents & LLMs
Ask My Portfolio
Three live demos of the agentic AI and LLM work Jose does: a retrieval-augmented assistant, a tool-using agent that shows its work, and a résumé-grounded content generator.
Demo 1 · Retrieval-Augmented Chat
Grounded answers with cited sources
Embeddings + cosine retrieval find the most relevant facts from a curated knowledge base, then the model answers from that context and cites what it used.
Ask anything about Jose's experience, skills, or projects. Answers are grounded in his resume and cite their sources.
Conversations may be logged to improve this demo — please don't share sensitive information.
AI assistant — it answers about Jose from his portfolio, doesn't speak for him, and can make mistakes.
Demo 2 · Tool-Using Agent
An agent that decides and shows its work
Given a question, the agent chooses which tools to call — portfolio search, a calculator, today's date — chains them, and synthesizes a grounded answer. Every tool call is shown so you can see how it reasoned.
Ask a question and watch the agent decide which tools to call — portfolio search, a calculator, and today's date — then chain them into a grounded answer.
Each tool call is shown as it happens.
Demo 3 · Content Generator
Tailored outreach, grounded in the résumé
Paste a role and pick a format and tone — the generator retrieves the most relevant résumé facts and writes a tailored pitch from them. It uses only what's in the résumé, so it never invents experience or numbers.
Three live demos — all run on a local Ollama model in development and OpenAI in production, and may occasionally be imperfect. The agent's calculator is a safe, arithmetic-only evaluator, and generated drafts are grounded in Jose's résumé.