Academic thesis review is a painstaking process. Manually checking 20+ pages for factual consistency and adherence to university standards is prone to human error.

The objective was to automate the verification of the diploma thesis against WZ UW requirements, legal frameworks like MiCA, and general factual accuracy to ensure a high-quality final submission.

Built an n8n automation workflow that extracts content from LaTeX files. I leveraged advanced LLMs for deep semantic analysis, integrated Tavily for real-time fact-checking, and utilized Pinecone vector database for RAG-based compliance verification.

Successfully reduced the analysis time to under 5 minutes, providing a detailed report that flagged exact locations of errors or inconsistencies. This allowed for rapid iteration and significantly higher document quality.