SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Marshmallow is a fintech unicorn (founded 2017, 700+ employees, £140M raised, profitable) on a mission to make migration easy. They've built insurance and lending products serving millions of customers and are scaling toward becoming one of the world's largest financial services providers.
You'll join the Data Science team as Principal Data Scientist in Claims, providing technical leadership across traditional ML and Generative AI systems. Claims is a critical customer touchpoint, and this role focuses on automating the claims journey through robust, production-grade decisioning systems.
Key responsibilities:
- Provide technical leadership for data science across Claims Fraud, shaping risk decisioning and fraud detection in partnership with Product and Engineering
- Design, build, and iterate on production ML and Generative AI/LLM systems supporting claims validation and automation
- Collaborate with other Claims data scientists to bring system-level thinking to how models, data, and workflows integrate; identify architectural improvements to scale decisioning and reduce time-to-production
- Advocate for platform and tooling investments (monitoring, feedback loops, QA) needed to achieve AI-driven end-to-end claims automation
- Champion robust, scalable, strategically aligned technical solutions in cross-functional discussions, ensuring systems and infrastructure support the multi-year vision for automated claims
- Set high standards for statistical rigor, experimentation, and measurement; improve how Claims performance and uncertainty are understood and communicated to senior stakeholders
You think in systems, connecting data science, engineering, and product to shape scalable solutions. You're confident challenging assumptions and influencing stakeholders across seniority levels with clear, pragmatic reasoning. You thrive in ambiguity, bringing structure and momentum to complex problems. You're motivated by real-world impact and driving meaningful automation and better customer outcomes.
The role is hybrid (3 days in office, London). Visa sponsorship available.
Requirements:
- Significant commercial experience delivering end-to-end Machine Learning solutions, from problem framing and experimentation through production deployment and ongoing monitoring
- Hands-on experience building and shipping Generative AI systems in production (not prototypes), including evaluation, safety/quality considerations, and integration into customer or operational workflows
- Strong statistical and modeling foundation with experience in risk-based decisioning under uncertainty (e.g., fraud, credit, insurance, or other regulated domains)
- Proven ability to influence technical direction across Data Science and Engineering, including shaping scalable model/service integration patterns and challenging proposals to drive robust, long-term solutions
- Strong stakeholder management skills with confidence communicating trade-offs and pushing back constructively with Product and Engineering to ensure high-quality outcomes
