What you’ll be responsible for

• Owning the development and continuous improvement of our production pricing models.
• Turning pricing opportunities into measurable hypotheses, experiments and production changes.
• Evaluating pricing performance through backtesting, online testing and commercial metrics.
• Developing Bayesian, multi-armed bandit and other optimisation approaches for dynamic pricing.
• Working with commercial, finance and product teams to understand pricing performance and recommend action.
• Building supporting risk models and agentic systems where they improve pricing, refund decision-making or operational efficiency.
• Applying rigorous offline and online evaluation to everything you build.
• Shipping reliable models through reproducible training pipelines, versioned artefacts, scheduled jobs, monitoring and alerting.
• Raising the team’s technical standards through code and model reviews, mentoring and the development of reusable patterns.
• Helping the Head of Data Science shape the team’s strategy and identify where investment in machine learning will have the greatest impact.
• Communicating findings, trade-offs and recommendations clearly to both technical and non-technical audiences.

What you’ll bring

• At least five years’ experience in data science or machine learning, including responsibility for models running in production.
• Experience of insurance, protection, travel, fintech, risk modelling or another transaction-led industry.
• Strong hands-on experience building pricing, revenue optimisation or commercially focused decision models.
• Expert Python skills, including production-quality code, testing and code review, alongside strong SQL.
• A solid grounding in machine learning techniques such as gradient boosting, GLMs, Bayesian methods and multi-armed bandits, or comparable experimentation and optimisation methods.
• Experience designing, running and evaluating pricing experiments using both commercial and statistical measures.
• The ability to connect model performance with commercial outcomes such as revenue, conversion and risk.
• Hands-on experience designing, building and optimising agentic systems, particularly graph-based and iterative workflows involving loops, branching, routing, tool use and state management.
• Experience evaluating LLM applications using approaches such as gold datasets, regression testing, LLM-as-judge scoring and canary testing.
• Experience with cloud ML platforms—ideally Azure ML, Functions and Blob Storage—or the willingness to transfer your knowledge to Azure.
• Experience with Git-based workflows, CI/CD and modern agent-assisted development tools such as Claude Code or Codex.
• The ability to turn an ambiguous business problem into a deployed, monitored solution.
• Strong commercial judgement and confidence discussing model decisions in terms of revenue and risk.
• Clear written and verbal communication, including with non-technical audiences.
• An evidence-led approach and experience mentoring others or providing technical leadership.
• Awareness of GDPR and international data regulations.

Useful, but not essential

• Causal inference and large-scale experiment design.
• MLOps, including model registries, artefact versioning and drift monitoring.
• Streamlit or similar tools for internal applications and dashboards.
• Experience with claims, refunds or customer operations.

You don’t need to match every item in this section. If the role sounds like a strong fit, we’d still like to hear from you.