Zühlke is developing a proof of concept (PoC) utilizing a large language model (LLM) to assess the viability of an existing text dataset. The goal is to determine if this dataset can effectively support a retrieval-augmented generation (RAG) system. This innovative project aims to structure complex data, facilitate accurate question answering, and provide grounded explanations for the responses generated. |
Skills: RAG architecture, Python, graph-based models Experience with CV (Curriculum Vitae)/JD (Job description) parsing, ontology integration, and evaluation metrics Familiarity with ESCO ontology and Databricks environment |
Software Development of PoC (Python, Databricks) RAG development, testing and fine-tuning Presentation and Visualisation of Results, Learnings and Challenges |

Data Scientist
Romania
Apply by 26 Nov 2025
40.0 - Per Hour
Job Ref.: 55439