Data Scientist Large Language Models

About VYDA

VYDA is a leading data expert in the German-speaking region, combining human expertise with innovative technology. As a reliable partner for process automation, VYDA supports data-driven transformation in the insurance and real estate sectors. The company, headquartered in Münster, has over 400 employees and merges deep industry knowledge with ISO-certified technology and years of experience in end-to-end business process management. VYDA strengthens the competitiveness of companies in auto, property, casualty, and real estate insurance with tailored solutions. AI-supported processes and professionally validated analyses provide the foundation for data-driven decision-making in insurance and real estate. VYDA collaborates closely with its strategic partner, the STIBL expert office in Austria.

Position Overview

Join VYDA GmbH as a Data Scientist focusing on Large Language Models (LLMs). This role offers ongoing challenges with significant autonomy for ideas and responsibility. Employees benefit from 30 fixed vacation days annually and flexible working hours, with the option to work entirely from home.

Key Responsibilities

• Develop machine learning algorithms including data extraction, preprocessing, feature engineering, model training, evaluation, and optimization.
• Design and enhance applications based on large language models by preparing and enriching text data, prompt optimization, and assessing model performance for accuracy, consistency, and reliability.
• Perform fine-tuning of LLMs to tailor models for domain-specific use cases, selecting appropriate training methods and datasets.
• Design and implement agent-based workflows in which LLMs autonomously plan tasks, integrate tools, and orchestrate multi-step processes.
• Develop Retrieval-Augmented Generation (RAG) systems involving vector database construction and optimization, retrieval pipelines, and integration of external knowledge sources.
• Work on various natural language processing topics, including text model creation, topic modeling, and transformer-based models.
• Standardize data and assist in identifying patterns within datasets.
• Support deployment, maintenance, and monitoring of AI models.

Candidate Profile

• Educational background: Completed degree in computer science or equivalent qualification.
• Professional experience: Proven expertise in data science, especially in natural language processing. Ideally, experience with LLM fine-tuning, agent-based workflows, or RAG architectures.
• Technical skills: Strong knowledge of Python, NLP, and deep learning. Familiarity with popular LLM frameworks such as LangChain, LlamaIndex, and Hugging Face, and fine-tuning techniques like LoRA and PEFT is advantageous.
• Language: Fluent in written and spoken German.
• Soft skills: Strong analytical thinking and conceptual abilities.
• Additional assets: Experience with SQL, transformer architectures, MLOps, Docker, vector databases (e.g., Pinecone, Weaviate, FAISS), and building agent-based systems is a plus.

Equal Opportunity and Contact

VYDA values diversity and encourages applications regardless of age, social background, gender identity, sexual orientation, nationality, or religion. The emphasis is on professional qualifications and personality fit. For any questions, Vanessa and Regina from the career team are available by email.