Senior Data Scientist

Description

You will be working for one of the largest automotive services businesses in the UK and Europe, providing end-to-end solutions across vehicle remarketing, fleet management, financial services and dealer software. Our Data Solutions team sits at the heart of this – building the intelligence layer that powers decisions across our business and for our  clients.

This is a senior individual contributor role within our Data Solutions team, reporting to the Lead Data Scientist. You will be a key part of our ambition to build a world-class data intelligence capability.

Your primary focus is the Decision Engine – a recommendation product we are building for major clients. The Decision Engine helps remarketing managers assess, price, channel and act on hundreds of vehicles in a single working session. Getting the underlying models right – and proving they work – is the difference between a product that gets used and one that gets ignored.

Beyond the Decision Engine, you will contribute to a growing intelligence capability that serves multiple business units across Europe. This is not a role where you disappear into a notebook for six months. You will work closely with product managers, domain experts and engineers, and you will be expected to make your work land.

Requirements

• Strong data science fundamentals – production-grade experience in machine learning, statistical modelling, time-series analysis or pricing/propensity modelling
• Proven track record of building models that go into production and stay there – not just analysis, but deployable, monitored, maintainable outputs that a product team can rely on
• Experience working in or very close to a commercial domain – you understand how businesses make decisions from data, not just how models work in isolation
• Comfortable working independently at pace – this role does not come with a lot of hand-holding; you need to be able to take a problem, break it down, and deliver
• Strong communication skills – you can explain model outputs and their limitations to product managers, engineers and senior stakeholders without hiding behind jargon
• Experience in automotive, vehicle remarketing, fleet management or fleet disposal – you understand how vehicles are valued, aged, channelled and priced, and you can apply that context to model design without needing to be briefed from scratch
• Familiarity with explainable AI techniques – SHAP, LIME, feature importance frameworks -and an instinct for when and how to apply them in a user-facing product
• MLOps experience – experiment tracking (MLflow, Weights & Biases), model monitoring, drift detection, automated retraining pipelines
• Experience with Databricks or a comparable lakehouse platform
• Python or R proficiency across the standard DS stack – pandas, scikit-learn, PyTorch or TensorFlow where relevant
• Experience with PySpark desirable

Job responsibilities

What You Will Work On

• Decision Engine models
  Building and owning the core models that underpin the Decision Engine – pricing intelligence, stock segmentation, buyer behaviour profiling, and channel optimisation. These models need to be accurate, explainable, and deployable into a production environment used daily by
  remarketing managers across multiple European markets.
• Counterfactual measurement
  Designing the framework that proves the Decision Engine works – what would have happened to vehicles that weren’t acted on by DE recommendations? This is one of the most important and most technically interesting problems on the roadmap. Without it, we cannot credibly demonstrate commercial impact to clients or to ourselves.
• Model explainability
  Building the trust layer. Remarketing managers will not act on recommendations they don’t understand. You will design explainability outputs – reason codes, confidence indicators, contributing factors – that make model outputs interpretable by non-technical users in a high- pressure working environment.
• Pricing intelligence
  Working alongside our Vehicle Data and Pricing capability to model  pricing decisions that account for residual values, market conditions, buyer segmentation and commercial policy. This is a complex, high-stakes domain – the kind of problem that rewards deep thinking over fast iteration.
• Raising the team’s ceiling
  You will be a senior voice in a team of five. You will share knowledge, review others’ work, and help build a culture where the team learns faster collectively than any individual could alone.