I'm a researcher working on AI for optimisation and decision-making, based in Porto, Portugal. My work sits between machine learning, operations research and engineering. I build methods that help systems make good decisions with limited data and expensive evaluations.
- Bayesian and multi-fidelity optimisation
- Safe, data-efficient learning for robotics
- Machine learning for logistics and last-mile delivery
- Operations research: routing, districting, mathematical programming
- High-performance and GPU computing for optimisation
| Project | Summary |
|---|---|
| Multi-Fidelity BO for Process Design | Using cheap and expensive simulators together to optimise chemical processes |
| Safe Learning for Robotic Assembly | Data-efficient controller tuning under safety constraints |
| Deep Learning for Last-Mile Delivery | Learning district visiting sequences and forecasting delivery demand |
| Last-Mile Planning Toolkit | Districting, distance prediction and route planning with an interactive interface |
Languages: Python, C++, Scala, SQL, JavaScript
ML and optimisation: PyTorch, scikit-learn, XGBoost, Gaussian processes, Bayesian optimisation, HiGHS, CPLEX
Data and compute: Apache Spark, pandas, GeoPandas, OpenMP/MPI, CUDA/CuPy
Tools: Git, Docker, Streamlit, Jupyter
- Hierarchical Temporal Convolution for District Sequencing in Large-Scale Last-Mile Delivery, OL2A 2024.
Most of my research code is private. Project pages give an overview of each piece of work.