University of Chicago PhD Student Riki Otaki Receives MongoDB PhD Fellowship Award
The University of Chicago Department of Computer Science is proud to announce that fourth-year PhD student Riki Otaki has been selected as one of the MongoDB PhD Fellowship recipients, recognizing his work on making large-scale data systems more efficient and scalable. The fellowship supports exceptional PhD candidates who are poised to make significant contributions to the future of computer science and related fields, and includes an award of up to $30,000 to support their research and training.
Otaki, a doctoral student mentored by Associate Professor Aaron Elmore, traces his interest in large-scale systems to a fascination with the challenges behind data-intensive science. He recalls being captivated by the Event Horizon Telescope project and the problem of processing petabytes of telescope data to reconstruct the first image of a black hole. What stayed with him was less the astrophysics and more the systems problem: taking huge, messy datasets and making them usable for computation.
When he looked for a place to pursue that kind of work, Elmore’s research group at UChicago stood out for its focus on resource-efficient system design and a strong culture of connecting ideas to real implementations. That environment has shaped his approach to data systems research, blending rigorous algorithmic thinking with attention to real-world workloads and constraints.
Otaki’s recent research centers on a fundamental challenge in modern data systems: indirection, the extra work required to find where a data record lives in memory, often by chasing pointers or consulting metadata structures. While such indirection is essential for flexibility and correctness, it can become a major bottleneck, adding CPU overhead and increasing contention as systems scale. His work develops a general approach to reducing these indirection costs during data access while preserving correctness.
Currently, Otaki is focused specifically on external sorting, the concept of sorting datasets that are far larger than main memory. His latest project studies how to efficiently sort datasets well over 100 GiB using under 5 GiB of memory, a regime that mirrors the constraints many systems face in cloud and multi-tenant environments where hardware resources are highly contested. He uses detailed benchmarking, which reveals where time and resources are actually spent in modern execution engines, with new system designs guided by those measurements.
Looking ahead, Otaki aims to continue pursuing systems research that bridges algorithmic ideas and production-relevant engineering—work that makes foundational data processing faster, cheaper, and easier to scale. He is especially motivated by building infrastructure that shortens the time from “we collected the data” to “we learned something,” making large-scale analyses and scientific results easier to produce and reproduce.
To have impact beyond a single prototype, Otaki focuses on techniques that can inform the design of real database engines and data platforms. This emphasis aligns closely with MongoDB’s interest in research that shapes the future of data-intensive applications and infrastructure. The MongoDB PhD Fellowship will support him in pushing these ideas further as he progresses toward completing his PhD.
“Receiving the MongoDB PhD Fellowship is a huge honor,” Otaki says. “It’s meaningful both as recognition of the direction of my research and as support that gives me more flexibility to pursue ambitious system-building and evaluation work, for example running large experiments that require significant compute time and storage. I’m grateful for MongoDB’s investment in academic research and excited to contribute to the broader database community.”
