Open to opportunities

Joao Manuel Borges Santos Da Costa

@joaomanuelborgessant

Senior data scientist and machine learning engineer specializing in time-series forecasting and production-ready predictive systems.

Portugal

Message

What I'm looking for

I’m looking for a team where I can build production forecasting systems with rigorous math and pragmatic engineering—especially time-series and energy/operations use cases—owning end-to-end analytics, improving accuracy (e.g., WMAPE), and deploying real-time solutions.

I’m a senior data scientist with 5+ years of experience building production-ready automated systems for forecasting, optimization, and predictive analysis. I bring a rigorous mathematical foundation, paired with pragmatic engineering, to simplify complex business problems into reliable, measurable outcomes.

I’m also a published researcher achieving SOTA results, including a 49% F1-score improvement on two-fold demand forecasting over previous SOTA architectures. I’ve deployed real-time systems for European-funded projects and delivered >20% improvement over major (+351) consulting firms.

Across my roles, I’ve built hourly probabilistic forecasting models for multiple commodities and focused on electricity demand models for hourly granularity forecasting—reaching < 3% WMAPE across the whole portfolio with hundreds of time series. I lead end-to-end analytics projects across demand forecasting, optimization, and customer segmentation, and I’m known for creating automated frameworks for feature creation, selection, and forecasting.

Experience

Work history, roles, and key accomplishments

Forecasting Expert

Sybilion

Dec 2025 - Jun 2026 (6 months)

Created automated frameworks for feature creation, selection, and forecasting pipelines, improving results by more than 20.5% for large enterprise clients. Developed scalable forecasting solutions for the AI & Data team across forecasting scenarios.

Feature Engineering Data Pipeline Development Performance Tuning Forecast Evaluation

Data Scientist / Analyst

Petrol Slovenia

Oct 2023 - Oct 2025 (2 years)

Led electricity demand forecasting for trading and operations, achieving under 3% WMAPE across the portfolio with hundreds of time series. Directed analytics projects spanning customer segmentation, demand forecasting, and optimization with hourly granularity.

Time Series WMAPE Evaluation Data Analytics Machine Learning Feature Engineering Model Monitoring

Senior Data Scientist

Jozef Stefan Institut

Dec 2019 - Oct 2023 (3 years 10 months)

Built production-ready automated forecasting and probabilistic hourly models for European-funded projects, achieving SOTA results including a 49% F1-score improvement on two-fold demand forecasting. Contributed peer-reviewed research focused on real-world time-series forecasting problems and deployed real-time systems.

Time Series Machine Learning Optimization Real Time Systems Feature Engineering Statistical Modeling Forecast Validation

Education

Degrees, certifications, and relevant coursework

University of Ljubljana

Master of Mathematics, Mathematics

2019 - 2023

Activities and societies: Focus areas included probability theory, statistical learning, stochastic processes, and information theory; coursework also covered LLM/NLP.

Master's in Mathematics at the University of Ljubljana covering probability theory, statistical learning, stochastic processes, and information theory.