Michael Dorner

1. High-Performance PDF Text Extraction for Academic Document Processing
   Develop scalable and layout-aware algorithms for extracting and structuring text from complex scientific PDFs (multi-column, math-heavy, or scanned). Focus on performance optimization, robustness, and data quality evaluation.
2. Design and Optimization of a PostgreSQL-Based Search Engine for Scientific Literature
   Implement and benchmark a high-performance search system for scholarly metadata and full text. Combine keyword and vector (semantic) search, optimize indexing, and evaluate scalability and ranking quality.

Requirements

• Strong technical background is mandatory.
• Proficiency in Go and/or Python is required.
• Familiarity with distributed systems, databases, or natural language processing is a plus.
• All theses must be written in English.
• Conference participation is encouraged, and travel will be sponsored if your work is accepted.