VLDB Honors Jana Giceva for Future-Proof Data Systems

The award was presented at the 52nd International Conference on Very Large Data Bases (VLDB 2026), held in Boston, Massachusetts. It honors researchers who have achieved significant impact through a specific technical contribution made after completing their Ph.D.

A New Vision for Data Systems

The award recognizes Giceva’s visionary contributions to building data systems whose core software abstractions decouple logical intent from physical hardware microarchitectures, thereby advancing the flexibility, extensibility and performance of data systems across evolving workloads and hardware platforms.

As data-intensive applications become increasingly central to science, industry and society, modern data systems face growing pressure to evolve. Traditionally, this has meant either building highly specialized systems that are difficult to reuse or continuously adding features to existing systems, making them increasingly complex and harder to maintain. Giceva’s research offers an alternative approach. 

Flexibility Without Performance Tradeoffs

A central idea behind the award-winning work is reducing the tight coupling between software components. Rather than binding, for example, a data format to a specific database engine or a computation to a particular processor, Giceva and her collaborators develop abstractions that allow these components to evolve independently while still exposing sufficient information for the system to optimize performance.

“The goal is to design systems that can adapt to change without compromising performance” says Giceva. “By rethinking some of the core interfaces and building blocks, we can create data systems that remain efficient over much longer periods of time.”

Lasting Impact on Data Systems Research and Design

The influence of Giceva’s research is already visible in both academia and industry. Her work on declarative sub-operators, presented at VLDB 2023, has inspired new directions in database system design. A notable industry example is Microsoft’s CoddSpeed framework for Microsoft Fabric: It adopts a similar strategy of decomposing complex database operators into smaller, reusable components, known as sub-operators, making it easier to map them to a variety of compute and network accelerators.

By demonstrating how data systems can remain both adaptable and efficient over time, Jana Giceva’s research is helping shape the next generation of data infrastructure. The VLDB Early Career Research Contribution Award 2026 recognizes a research vision that is already influencing both database research and industrial data system design.