Marcin Żukowski receives CWI Dijkstra Fellowship for pioneering work in database systems

Centrum Wiskunde & Informatica (CWI) awards the Dijkstra Fellowship on 21 November to Marcin Żukowski, former CWI database researcher and co-founder of the globally successful tech company Snowflake. Żukowski's pioneering work produced techniques that are still essential for efficiently processing huge amounts of data, leading to faster analysis and new opportunities for companies working with big data.

Publication date
11 Nov 2024

With the Dijkstra Fellowship, named after Dutch computer pioneer Edsger W. Dijkstra, CWI honours Żukowski's exceptional scientific and technological contributions, which have led to applications with major societal impact. For example, Żukowski's innovations have now been integrated into technologies that drive Snowflake's global success and play an important role in database management combined with cloud solutions.

"Marcin is an excellent example of how CWI's mission can be put into practice. He used his PhD research at CWI to create versatile fundamental software products that are now widely used, sharing his knowledge and experience with the public and in particular with young technology entrepreneurs," said CWI director Ton de Kok.

Portrait of Marcin Żukowski
Marcin Żukowski at the Dijkstra Fellowship & Lectures. Picture: Kevin Knipping

Pioneer in database technologies

Marcin Żukowski began his career at CWI, where he completed his PhD research on database management system architectures. There, he developed the innovative concept of vectorised execution: here, a computer processes data in groups (vectors) instead of one by one. This speeds up the process of searching and analyzing information in large data sets, because the same operation can be performed for several data simultaneously. This research earned several scientific awards, including the CIDR 2024 Test of Time Award, established by the Conference on Innovative Data Systems Research.

After his PhD, Żukowski founded the CWI spin-off VectorWise (now Actian) with fellow researchers from the CWI Database Architectures group, where he translated his research into a fast analytical database system with high scalability. The spin-off experienced rapid technological and commercial growth: VectorWise became the blueprint for analytical databases and it is still a widely used database system.

Data storage in the cloud

In 2012, Żukowski left the spin-off with destination Silicon Valley when Thierry Cruanes and Benoit Dageville asked him to become Snowflake's third founder. The aim was to get the technologies Żukowski had invented - in addition to vectorised execution, new lightweight compression methods - into the Snowflake architecture. Żukowski built up and led the tech company's development department in its early years.

Snowflake was the first database system specifically designed to store data externally, in the cloud. This means, among other things, that the calculations required to process database queries (searches) are completely decoupled from storage. In this way, cloud services become scalable, so that when the workload increases, more computers can quickly be switched on without having to copy and redistribute the data first.

Earlier this year, Żukowski left Snowflake. He is currently active as an investor and consultant. He also supports technology development and innovation in his homeland Poland.

Picture at very top: Kevin Knipping

Portrait of Edsger Dijkstra

About the Dijkstra Fellowship and the Lectures

The Dijkstra Fellowship is named after former CWI researcher Edsger W. Dijkstra, one of the most influential scientists in CWI's history. Among other things, Dijkstra developed the shortest path algorithm.


About the Dijkstra Fellowship Database event
Żukowski will give a special lecture on 21 November in which he will share his experience in developing academic technologies into successful products. In addition, peers will give lectures on the architecture of systems for data processing and analysis.

Picture: CWI/Annette Kik