At TÜV SÜD we are passionate about technology. Innovations impact our daily lives in countless ways, and we are dedicated to being a part of that progress. We test, we audit, we inspect, we advise. We never stop challenging ourselves for the safety of society and its people. We breathe technology, we strive for professional excellence, and we leave a mark. We take the future into our hands. We are TÜV SÜD.

Tasks

• Design and implement production-grade pipelines for extracting structured information from unstructured documents using a mix of LLMs, machine learning models, and rules-based logic
• Contribute to the improvement and extension of an established, AI based platform for document analytics, focusing on robustness, scalability, and cross-use-case adaptability and support the continuous evaluation, and refinement of deployed models to maintain accuracy and transparency
• Conduct applied research and solution ideation on complex document analytics tasks, such as information extraction, classification, entity recognition, and semantic matching
• Design and prototype methods using both traditional NLP and large language models (LLMs) to identify and link information in technical documents to general regulatory requirements
• Collaborate with data engineers, solution owners, and subject matter experts to ensure technical feasibility, domain relevance, and alignment with operational constraints
• Communicate technical decisions, trade-offs, and constraints clearly to non-technical stakeholders and management; contribute to the alignment of the technical roadmap with product and business priorities.

Qualifications

• Degree in Computer Science or equivalent, with a minimum of five years of professional experience as a Data Scientist, with a focus on NLP and document analytics and proven track record in developing and deploying solutions for text and document analysis, including the practical use of large language models (LLMs)
• Experience with emerging LLM-based methods, including agentic workflows or multi-step reasoning approaches for complex document understanding tasks
• Demonstrated success in implementing analytical models as production-ready components, integrated into business-critical applications
• Strong programming skills in Python, with deep familiarity across the data science and NLP ecosystem (e.g., pandas, scikit-learn, spaCy, Hugging Face, LangChain) and practical experience with Microsoft Azure, including Azure DevOps and Azure OpenAI services
• Comfortable working in agile, sprint-based teams in cross-functional, interdisciplinary environments
• Excellent communication skills in German and English

At TÜV SÜD, we have employees from more than 100 different countries collaborating together. People of different backgrounds, skills, and pursuing different life goals. Our strength comes from these countless and varied perspectives.

We are committed to be an inclusive and diverse workplace by welcoming people of all backgrounds. We want Diversity & Inclusion (D&I) to be a foundation of our company and create an environment where all our employees can trust they will be treated with respect, regardless of gender, nationality, ethnic background, faith, beliefs, disabilities, age, sexual orientation, or identity. As such, our employees are expected to behave at all times in a manner consistent with TÜV SÜD Code of Ethics and Company values.

We firmly believe embedding D&I in the heart of what we do will inherently contribute to the success of TÜV SÜD. Click here to find out more about Diversity at TÜV SÜD.