I’m Sebastião Santos Lessa, based in Porto, Portugal 🇵🇹.
Professionally, I work as a Generative AI Engineer, building LLM based systems like conversational agents 💬, multi agent workflows 🧠🧩, and AI powered automation ⚙️🤖. A big part of my work lives in the world of Retrieval Augmented Generation (RAG) 🔎📚, from knowledge ingestion and embeddings to retrieval pipelines, prompt engineering, and evaluating grounded, reliable outputs in production.
Alongside GenAI, I also work on data science and machine learning projects 📊📈, covering the full journey from data exploration and modelling to deployment and monitoring in real systems.
From a scientific perspective, my academic focus is on time series analysis and forecasting ⏱️📉, which is also the core topic of my Master’s thesis. I’m especially interested in foundation models for time series, synthetic data generation, imputation, and modelling complex temporal behaviour in real world systems, with a strong focus on energy related domains ⚡.
I enjoy working where theory meets practice 🧠➡️🌍, using rigorous methods while keeping a clear focus on applicability and measurable impact.
I am the author of a peer reviewed research paper, presented at IEEE PowerTech 2025 ⚡🎓, focused on the use of foundation models for time series imputation in energy systems.
Paper link 👉 IEEE
🎓 BSc in Artificial Intelligence and Data Science
Joint degree FCUP × FEUP · University of Porto
🦾 MSc in Artificial Intelligence
Joint degree FCUP × FEUP · University of Porto
🖥️ For my command-line interface (CLI), I prefer using the hyper terminal with the zsh shell. It offers a great user experience and makes my development workflow more efficient. With the PowerLevel10k theme.
| Configurations: | hyper.js | zshrc | p10k.zsh |
|---|