Congratulations to Manuel Widmoser on successfully defending his dissertation

On October 31, 2025, Manuel Widmoser successfully defended his cumulative Ph.D. thesis “Scaling Similarity Queries to Massive Datasets”.

Thereby, Mr. Widmoser completed his doctoral studies and received the title “Dr. techn.” (equiv. Ph.D.).

Over the last four years, Mr. Widmoser diligently worked on multiple similarity query techniques and how to efficiently execute them on very large datasets. As part of his Ph.D. thesis, Mr. Widmoser introduced the following novel techniques: (1) A robust similarity join algorithm that leverages metric properties to provide fast runtimes and low memory consumption over a wide range of different dataset characteristics. (2) The first design for distributed inverted list index under memory disaggregation (the current architectural trend for large data centers). Ultimately, this enables efficient similarity query processing in a distributed setting, i.e., if the index no longer fits into the main memory of a single machine. (3) The first graph-preserving design of the Hierarchical Navigable Small Worlds (HNSW) index under memory disaggregation that scales beyond billions of vectors while retaining near in-memory performance and high accuracy.

After Mr. Widmoser’s successful defense, we celebrated this milestone with family, friends, and colleagues.

About Manuel Widmoser:

Manuel Widmoser earned his Bachelor’s and Master’s degree in Computer Science at the University of Salzburg. In 2021, he started his doctoral studies as part of the database research group at the Department of Computer Science, supervised by Prof. Nikolaus Augsten and Daniel Kocher (Co-Advisor).