Skip to content

Projects

Applied machine learning for urban transport and city logistics: from probabilistic forecasting to live services and operational decisions.

Forecasting shared mobility in the RET app

2023–2026
Forecasting shared mobility in the RET app illustration

SUM · Industry deployment

I led modelling and coordinated delivery of a live service forecasting CHECK vehicle availability in Rotterdam, working with RET, Siemens Mobility and Infoplaza. Read the case study.

Live serviceXGBoostIndustry collaboration

T-STAR

Published 2026
T-STAR illustration

Probabilistic demand forecasting

A two-stage transformer framework for sparse demand at 15-minute resolution, evaluated across 235 bike-sharing stations, including transfer to unseen service areas. Read the case study.

TransformersTime seriesOpen source

STAMP

Under review
STAMP illustration

Joint demand and usable-supply forecasting

An adaptive graph-based multi-task framework jointly forecasting demand, fleet size and low-battery vehicles up to four hours ahead. Read the case study.

Graph learningMulti-task learningUncertainty