I’m Sebastião Santos Lessa, a Generative AI and Machine Learning Engineer based in Porto, Portugal 🇵🇹.

I build GenAI systems that connect LLMs, backend services, cloud infrastructure, and enterprise workflows. My work includes conversational agents, agentic workflows, tool calling, Retrieval Augmented Generation, and AI powered automation.

I am particularly interested in how AI agents reason, use tools, manage context, and execute complex workflows.

Alongside GenAI, I work on machine learning and data science projects. My public projects cover areas such as time series forecasting, reinforcement learning, and applied machine learning.

My academic work focuses on time series analysis and forecasting, particularly foundation models for time series, anomaly detection, synthetic data generation, imputation, and energy systems.

I enjoy working where research meets engineering, combining experimentation with practical applications.

I authored a peer-reviewed paper presented at IEEE PowerTech 2025, focused on foundation models for time series imputation in energy systems.

Read the paper on IEEE 📄

My master’s thesis focuses on adaptive anomaly detection in energy consumption time series, using probabilistic forecasting to account for changing consumption patterns.

View the thesis at U.Porto 📄