Wilson Llanos

AI Engineer

I build applied AI systems, from prototype to production.

Download CV

Call me: +55 11 97522-5763

Portrait photo of Wilson Llanos, studio shot

Summary

I am a Systems Developer at EISA, building production generative AI solutions on the Azure ecosystem with a focus on data security and governance. I built a multi-agent system with LangGraph and MCP for IT incident triage, reducing response time, and I am now building a multi-agent system with hybrid RAG to automate tier-1 support. I also work on data integration across four systems, including legacy ones, eliminating recurring errors, and I have Machine Learning projects with Apache Spark on Databricks. MBA in AI and Big Data from ICMC/USP (completion in October 2026) and a degree in Civil Engineering.

Timeline

  1. 2026-01 — Present · 💼

    Systems Developer · EISA - Empresa Interagrícola

    Promoted after exceeding goals, I brought RAG and LLMs to the team: I am building a multi-agent system to automate tier-1 support, aiming to reduce ticket volume and the IT team’s response time.

  2. 2025-07 — 2026-10 · 🎓

    MBA in Artificial Intelligence and Big Data · ICMC / USP

    A strategic decision to deepen my AI education: machine learning, deep learning and generative AI, plus Big Data. Expected completion in October 2026.

  3. 2025-02 — 2025-12 · 💼

    Programmer Analyst · EISA - Empresa Interagrícola

    First promotion, earned by delivering under demanding deadlines. I advanced in data integration, with ETL pipelines and analytical reports for the coffee business, and started my first generative AI prototypes with RAG and LLMs.

  4. 2025-01 · 🎓

    ETL Pipelines and Machine Learning with Apache Spark · Data Science Academy

    My starting point in the AI world: ETL pipelines and Machine Learning models with Apache Spark.

  5. 2022-04 — 2025-01 · 💼

    Support Analyst · EISA - Empresa Interagrícola

    My start in IT: supporting users across systems, databases and infrastructure, reducing recurring tickets, while building up SQL, C#, CI/CD and Azure skills for the move into development.

  6. 2012-01 — 2016-12 · 🎓

    Bachelor's in Civil Engineering · Uninove

    The math and statistics foundation that eased my move into technology and still supports my AI studies.

Scroll to see more ↓

See full details in my résumé (PDF)

Projects

Personal · open source

Portfolio Portal with AI Chat

This website: a bilingual portfolio with an AI assistant that answers questions about my career using RAG over my own résumé, with a prompt-injection guardrail, cost limits and CI/CD. Built with Spec-Driven Development.

  • Next.js
  • TypeScript
  • FastAPI
  • Gemini
  • RAG
  • Prompt Guard 2
  • Upstash Redis
  • GitHub Actions
  • Vercel
View on GitHub →

Technical challenge (Getnet case) · open source

Multi-Agent Merchant Support

A LangGraph router sends each message to specialized agents: RAG over approved sources, the merchant’s own data via tools, and external APIs, with guardrails, personal-data redaction and observability. When context is insufficient, it admits the limitation instead of making things up.

  • Python
  • FastAPI
  • LangGraph
  • LangChain
  • Claude
  • Chroma
  • Sentence Transformers
  • Docker
View on GitHub →

Used at EISA · open source

Retrieval Quality Evaluation

A command-line tool that measures the search quality of a documentation corpus before publishing it to Azure AI Search: it compares vector and hybrid search (BM25 + RRF) on a golden set, with Recall@k and MRR against a baseline.

  • Python
  • Azure OpenAI
  • Qdrant
  • BM25
  • fastembed
  • Typer
View on GitHub →

Certifications