About me

I am a GenAI / ML engineer based in Guimarães, Portugal, with a long-running focus on NLP and systems people actually open. I ship production AI products — a Portuguese legal RAG assistant (LexPT) and a bilingual fitness coaching SaaS for a paying trainer — and maintain a personal knowledge stack: YouTube → LLM wiki → spaced-repetition apps, Chinese vocabulary SRS, and study tools.

Trilingual in Portuguese (native), English (fluent), and Mandarin (native). Outside of engineering I referee football for AF Braga, train streetlifting, run endurance events (Porto Marathon, Nov 2025), and compete in Go (national silver). Open to relocation across Europe, Asia, and the Middle East. Targeting GenAI / ML roles focused on RAG, LLM applications, and production delivery.

Selected work: LexPT · FitnessAPP · Streetlifting · Wiki Flashcards · Second Brain

What I'm doing

  • GenAI & RAG

    Production RAG pipelines with hybrid retrieval, reranking, and grounded multi-step generation (LexPT).

  • Full-stack apps

    Next.js, Firebase, FastAPI, and Docker — shipping products used in day-to-day coaching operations.

  • Knowledge systems

    Second-brain wiki pipelines, 1,184 flashcards, Chinese SRS, and study apps on GitHub Pages.

  • ML & NLP

    Forecasting, sentiment analytics, and a thesis on hierarchical time-series anomaly detection.

Looking for

  • Roles GenAI / Machine Learning Engineer (junior to mid) owning RAG pipelines, LLM product features, and full-stack delivery.
  • Domains LegalTech, coaching/health apps, telecom/ops anomaly detection, multilingual European teams.
  • Location Hybrid or on-site in Portugal (Porto / Braga / Lisbon), or relocation across Europe, Asia, and the Middle East.

Resume

Education

  1. University of Minho — M.S. EGSI

    Sep 2019 — Present

    Engineering & Management of Information Systems. Coursework complete; thesis in progress: hierarchical anomaly detection in multi-dimensional time series. Coursework: Machine Learning, Data Analysis, Big Data Systems, Statistical Modeling, COBIT / ITIL / McFarlan / SISP.

Experience

  1. Independent / Client Engineering

    2021 — Present

    Deliver and maintain Sergio Oliveira PT SaaS for a paying client; iterate LexPT and public knowledge apps. Operate YouTube → LLM wiki → flashcard sync (Second Brain) and publish study apps on GitHub Pages.

  2. AF Braga — Football Referee

    Jan 2021 — Present

    Officiate matches under pressure with clear communication and rule-based reasoning. Built exam Q&A NLP, flashcards, and reminder automation used by 50+ referees preparing for licensing exams.

Selected work

  1. LexPT — Portuguese Legal AI

    Jan 2024 — Present

    RAG over 5 tribunal corpora + EU legislation; hybrid ChromaDB + BM25 + reranking; six-step grounded LangChain generation; Gradio Q&A; FastAPI / Dify paths. GitHub

  2. Sergio Oliveira PT — Fitness SaaS

    Mar 2023 — Present

    Production gym platform (Next.js 15, Firebase, Genkit): bilingual portals, billing, AI workout drafts. Near-zero infra cost (replaced ~€150/month server). GitHub

Technical skills

  • GenAI & LLMs
    LangChainDeepSeekGeminiGenkitOllamaDifyGradio
  • RAG & retrieval
    ChromaDBBM25E5 / SerafimCross-encoder
  • Apps & backend
    PythonFastAPIFlaskNext.jsTypeScriptFirebaseDockerExpo
  • ML & data
    PyTorchTensorFlowScikit-learnpandasSQLRSparkPower BITableau
  • Also
    JavaC++GitHub ActionsViteObsidian pipelines
  • Spoken languages
    Portuguese (Native)English (Fluent)Mandarin (Native)

Certifications

  1. Deep Learning Specialization

    Coursera · DeepLearning.AI

    Course notebooks

  2. Machine Learning Specialization

    Coursera · DeepLearning.AI

    Course notebooks

  3. Master Microsoft Power BI

    LinkedIn Learning

    Dashboarding and BI storytelling.