Data Scientist

Freelance Data Scientist (Insurance) – AI Transformation

Location: Remote (Occasional travel to Munich)

Contract Length: 6–12 months (with extension)

Day Rate: €700–€850/day (depending on experience)

Language: Fluent German required

Project Overview

We are partnering with a leading Munich-based Insurance group who are looking for a Freelance Data Scientist (Insurance) for a major AI and data transformation programme. Leveraging Databricks and the Lakehouse architecture, this project is redefining how pricing, underwriting, and fraud detection are powered through real-time analytics and machine learning.

This is a high-impact, strategic initiative where you’ll help scale production-ready ML workflows, modernise legacy systems, and work alongside experienced data scientists, actuaries, and cloud engineers.

Key Responsibilities

• Design and deploy machine learning pipelines using Databricks (PySpark, MLflow, Delta Lake)
• Develop predictive risk models, pricing algorithms, and fraud detection systems using customer, claims, and behavioural data
• Implement NLP solutions to extract insights from unstructured claims documentation
• Optimise and monitor real-time data pipelines in collaboration with data engineering teams
• Support transition from legacy analytics tools (e.g., SAS) to a modern Python/Databricks stack
• Ensure full regulatory and GDPR compliance, including model transparency and auditability

Key Requirements

• 5+ years’ experience in Data Science, ideally within Insurance or Financial Services
• Proven hands-on experience with Databricks, PySpark, MLflow, and Delta Lake
• Strong coding skills in Python and SQL
• Experience with NLP, fraud analytics, and predictive modelling
• Familiarity with cloud platforms (Azure preferred, also open to AWS/GCP)
• Knowledge of claims workflows, underwriting, or actuarial models is a bonus
• Fluent German (B2/C1+) essential

If you are interested in the Freelance Data Scientist (Insurance) position, please apply with an up-to-date CV as soon as possible.