Machine Learning / Data Scientist

Twinspire is building the data and intelligence layer for elite football by connecting GPS, wearables, testing, recovery, and contextual data to learn how each athlete responds to training over time. We're now entering the next phase. As we prepare deployments with professional football clubs, we're looking for someone who wants to help build and validate the modelling layer behind our individual Training Response Profiles. Why is this problem interesting? The problem is interesting because the data is messy by nature:

• longitudinal and highly individual
• multimodal and asynchronous
• sparse and missing in non-random ways
• influenced by training dose, recovery, travel, context and adaptation
• collected from athletes rather than controlled laboratory systems

We're particularly interested in people with experience or strong interest in:

• time-series / longitudinal modelling
• hierarchical and personalised ML
• multimodal data fusion
• uncertainty and missing-data modelling
• sequential/temporal models
• process mining and event-based representations
• rigorous model validation and benchmarking
• Python and strong statistical/ML fundamentals are important.

Experience in sports science is absolutely not required. What matters much more is scientific curiosity, comfort with challenging assumptions, and a desire to build models that can survive contact with real-world data. Working with us You would work closely with the CEO, the engineering team, and researchers across DTU, with access to a growing longitudinal dataset and real deployment environments in professional football. The role can start part-time alongside an MSc/PhD and has the potential to develop into a full-time position as Twinspire grows. If this sounds like you - or someone exceptionally good comes to mind - send us a message!