Description:
We are looking for a Machine Learning Engineer to help build frontier models to understand and improve complex operational systems. The work sits at the intersection of scientific machine learning, time-series modelling, optimisation and real-world deployment. You will work closely with the founding team, customer sites and industrial data to turn early technical validation into a scalable product.
This is a hands-on engineering role. You will not just train models in isolation. You will build systems that need to work with messy data, operational constraints and real-world environments.
What you will do:
- Design, train and deploy machine learning models for complex operational systems.
- Work with sparse, noisy and irregular time-series data from real-world environments.
- Build models that combine data-driven learning with physical and operational constraints.
- Develop reusable modelling components that can scale across different sites and use cases.
- Work with the product and engineering team to move models from prototype to production.
- Evaluate model performance, reliability and robustness in applied settings.
- Spend time with customers to understand the operational context behind the data.
- Contribute to the technical direction of the platform as one of the first ML hire
What we are looking for:
- A degree in machine learning, computer science, engineering, physics, mathematics, applied mathematics, operations research or a closely related STEM field from a top university.
- Strong practical experience building machine learning models in Python, ideally using PyTorch, JAX or similar frameworks.
- Experience with one or more of: scientific machine learning, physics-informed ML, time-series modelling, optimisation, simulation, forecasting or probabilistic modelling.
- Comfort working with messy real-world data, including missing values, drift, noise and inconsistent data quality.
- Interest in applying machine learning to physical systems, industrial operations and real-world optimisation problems.
- In-person working from our London office, typically 4–5 days per week, with occasional travel to customer sites in the UK