Meewerkstage Data Science LLM Quality Assessment

This is a Data Science internship (meewerkstage) at a.s.r. focused on Large Language Model (LLM) Quality Assessment. As an intern within the Data Science & AI team, your primary objective will be to develop, implement, and refine evaluation frameworks to assess the quality, accuracy, safety, and reliability of LLMs deployed within the organization's insurance and financial services workflows. Daily responsibilities include benchmarking model performance, analyzing hallucination rates, engineering prompt templates, and investigating LLM-as-a-judge methodologies. You will leverage statistical modeling techniques, natural language processing (NLP) metrics (such as BLEU, ROUGE, and semantic similarity scores), and evaluation tools to ensure robust model alignment and quality standards. The required programming and technical stack includes Python, Hugging Face transformers, LangChain or LlamaIndex, PyTorch, and version control via Git. This position offers a hands-on learning experience in applying state-of-the-art Generative AI quality assurance techniques within a leading financial institution.