LON Data Scientist

LON Data Scientist

Working hours: This role is available on a part-time, job-share or full-time basis.

Salary: Up to £50,000 depending on experience plus an excellent benefits package

Location: Swindon, Fareham or London. Hybrid Working.

Closing date for applications: 30th January 2026

Data is central to our work at Zurich, and we seek a talented individual to join our data science team focusing on topics related to our Retail business. You will collaborate with data, AI and business experts to support data-led decisions, impacting the entire organization. Our data & AI journey is entering an exciting new stage, and you can help shape its future.

We aim to put people and their data at the heart of our strategy, enabling swift, outcome-focused decisions. You'll work with colleagues to design core systems and data requirements, bringing efficiency and supporting colleagues in using data confidently.

We seek someone passionate about AI, data and advanced analytics, willing to challenge the status quo, and proficient in AI/ML in a commercial setting. The ideal candidate is creative, curious, and logical in problem-solving. This is an excellent opportunity to join a dynamic, supportive team working on diverse insurance topics. We value a supportive team culture and independence in work. If this sounds like you, we look forward to your application.

We offer flexible working arrangements, including part-time, flexible hours, job share, remote work, or compressed hours, as we aim to attract the best talent. Discuss your flexibility needs with us during the interview.

• Driving data-led decision-making in our business, working with stakeholders to help them understand how data & AI can assist them in meeting their requirements. You will assist colleagues in setting strategy for activities using data & AI, informing on the art-of-the-possible and best-practice.
• The role-holder will use AI/ML and data science techniques to reduce the need for manual work, including analysis of structured numerical data and application of LLMs and similar tools to unstructured data. They will promote an automation-first mindset where possible and use data to tell compelling narratives and communicate in simple language to deliver tangible impact. This will involve using established AI technologies and exploring future AI capabilities.
• The role-holder will be a vocal proponent of both good data & AI practices and will be able to communicate these ideas to non-experts. They will ensure that high standards of code development are met, including adherence to code management best practices and team policies such as submission and review of pull requests. They will work with the business teams to build knowledge and confidence with data as well as collaborating with colleagues across the Zurich Group internationally to share knowledge and enhance the data analytics capabilities for our collective AI communities.

• Strong analytical, structured, and interdisciplinary way of thinking and working, including the ability to think creatively with data and being comfortable with complex and ambiguous problem-solving.
• Enthusiastic to work on problems which have never been attempted before.
• Proficient in Python and modern software development practices within a team of developers e.g. use of Git.
• Experience using SQL and working with databases. Comfortable working with a variety of data sources, both structured and unstructured and very large datasets using distributed computing (e.g. spark).
• Experience working with LLMs to deliver value in a commercial organisation, including how to manage and monitor LLM-based applications to maximise performance.
• Experience working with cloud technology, ideally Microsoft Azure and/or AWS.
• Proven track record of development and deployment of machine learning algorithms, including supervised and unsupervised learning techniques.
• Excellent collaboration and organisation skills.
• Proficient communicator, who is comfortable explaining the value of their work to drive adoption and challenge the status quo, both to technical and non-technical audiences.
• Comfortable…