Data Scientist (f/m/d)

For the past 10 years, omni:us has been pioneering AI for insurance. Long before AI became a boardroom priority, we were already applying it to one of the industry's hardest environments: complex, data-heavy, highly regulated insurance operations.

Today, our technology is used to understand claims, automate work, support complex decisions and run entire process steps at scale. We combine advanced AI, deep insurance expertise and production-grade engineering to turn what is technically possible into systems insurers can actually operate. We are not exploring what AI might mean for insurance. We have spent a decade building it.

Our focus now is on agentic and generative AI: AI agents that read and understand insurance documents, reason over policies and claim files, use tools and data sources, and either complete work on their own or assist human experts in their daily work.

As part of adesso, we take that experience further: combining omni:us' focus, speed and AI depth with the reach, implementation strength and enterprise access of one of Europe's leading IT service providers.

The next chapter of AI in insurance is already underway. At omni:us, you help build it from the inside.

We're looking for a Data Scientist to design, build and evaluate generative and agentic AI solutions for insurance automation and assistance. You'll turn LLMs and multimodal models into reliable systems: agents that handle claims and service workflows from end to end, and assistants that support claims handlers, underwriters and service teams with summaries, recommendations and answers grounded in the customer's data.

You will work at the point where cutting-edge AI meets real operational complexity. The technology you develop does not stay in a lab, a paper or a prototype. It goes into production, handles real insurance processes and creates measurable impact.

This role suits someone who cares as much about evaluation and reliability as about new ideas, and who wants to see their work make a real difference for clients.

• →Design and build LLM-based agents and assistants for insurance workflows such as claims intake and triage, coverage checks, document and correspondence processing, fraud indicators, and case summarisation.
• →Develop agentic workflows that combine LLMs with tools, APIs, business rules and retrieval over policies, claim files and knowledge bases (RAG), including multi-step reasoning and human-in-the-loop handover.
• →Apply and adapt multimodal models for extracting information from complex insurance documents (forms, invoices, reports, emails, images).
• →Build robust evaluation frameworks for generative and agentic systems: test sets, automated and LLM-as-a-judge evaluation, accuracy and hallucination metrics, cost and latency tracking, and monitoring in production.
• →Work on prompt and context engineering, and where it pays off, fine-tuning or distilling models for specific insurance tasks.
• →Design guardrails for safe, explainable and auditable AI decisions, in line with regulatory requirements such as GDPR and the EU AI Act.
• →Collaborate with product managers, engineers and insurance domain experts to bring solutions into production and measure their business impact.
• →Stay up to date with developments in generative and agentic AI, assess new models and frameworks, and bring the useful ones into our products.
• →Document approaches, experiments and results for internal use and for communication with clients.

• Bachelor or Master degree in Data Science, Machine Learning, Artificial Intelligence, Computer Science, or a related field.
• 3+ years of experience in data science or machine learning, with a proven track record of delivering AI solutions to production.
• Hands-on experience building applications with LLMs (e.g. RAG, tool use / function calling, structured outputs, agents).
• Experience in the insurance or financial industry is a plus but not required.
• Strong Python skills and experience with ML/DL libraries such as PyTorch, Hugging Face or scikit-learn.
• Experience with LLM APIs and agent frameworks or protocols (e.g. LangGraph, LlamaIndex, Model Context Protocol, or similar).
• Good understanding of how to evaluate and monitor generative AI systems, including dealing with hallucinations, non-determinism and cost.
• Familiarity with NLP, document understanding and/or multimodal models.
• Solid understanding of data engineering concepts (ETL, databases, vector stores).
• Experience with Linux (bash, Docker) and cloud platforms (AWS, Azure) is beneficial
• Proficiency in German and English (written and spoken).
• Ability to work with and analyze multilingual data effectively.
• Strong analytical and problem-solving abilities, with a pragmatic, product-oriented mindset.
• Excellent communication skills for working with technical teams, domain experts and clients.
• Self-driven and detail-oriented, with a healthy scepticism towards hype and a passion for what actually works.

If you're excited about building AI agents and assistants that transform the insurance industry, we'd love to hear from you. Please submit your resume and a cover letter highlighting your relevant experience to application link. Include any links to projects, GitHub profiles or portfolios that showcase your work, especially anything you've built with LLMs or agents.

omni:us is committed to fostering an inclusive and diverse workplace. We welcome candidates from all backgrounds and experiences to apply.