Staff Data Scientist – Claims ML & AI Systems

Marshmallow; Greater London; Hybrid; GBP 90,000 - 150,000; Full time

Benefits offered by this job

Bonus scheme

Private medical insurance

Learning budget and 2 dedicated L&D da

Monthly flexible benefits budget

25 days holiday

4 weeks Work From Anywhere

Job summary

Marshmallow in London is seeking a Staff Data Scientist to lead the Claims data science agenda, building robust ML and Generative AI systems that drive automated decisioning and smarter customer journeys.

You will collaborate across Product, Engineering and Operations to design scalable solutions, and raise the bar for statistical rigor and governance. This role offers hybrid work in London with real opportunities to shape high-impact AI initiatives.

Qualifications

• Experience delivering end-to-end ML solutions in production.
• Experience shipping Generative AI systems in production.
• Strong statistical foundation in uncertainty and risk-based decisioning.
• Experience influencing technical direction across teams.
• Strong stakeholder management and communication.

Responsibilities

• Lead data science for Claims Fraud and risk decisioning.
• Design and deploy production ML/GenAI systems for claims.
• Collaborate with DS and Eng to scale models and workflows.
• Advocate for robust platform investments (monitoring, QA).
• Drive measurable impact on automated claims processing.
• Set a high bar for rigor and experimentation.

Skills

End-to-end ML

Generative AI

Risk decisioning

Stakeholder mgmt

Production deployment

Tools

Python

Spark

ML Pipelines

Job description

Marshmallow in London is seeking a Staff Data Scientist to lead the Claims data science agenda, building robust ML and Generative AI systems that drive automated decisioning and smarter customer journeys.

You will collaborate across Product, Engineering and Operations to design scalable solutions, and raise the bar for statistical rigor and governance. This role offers hybrid work in London with real opportunities to shape high-impact AI initiatives.