PL/SE Seminar - Trustworthy Programming for Online Data Processing
Abstract: Today, data is generated with higher volume and higher velocity than can be feasibly stored, raising the need for new algorithms, software abstractions, and systems. However, online systems frequently fail, causing catastrophic outages, and leading to permanently dropped data and permanent loss of public trust. My research proposes programming abstractions and software tools with formal correctness guarantees for the online data processing paradigm. In part I of the talk, I discuss my PhD work on distributed streaming analytics platforms. We leverage partial orders for correctly composing and distributing data processing operators. My research provides a number of open source tools including DiffStream, which allows programmers to write ordering constraints and check for concurrency bugs at runtime. In part II, I discuss my work on developing core verification technology to be more efficient. My work is shipped with the Z3 Theorem Prover and has applications in monitoring security in the cloud. These directions come together in future work, where my goal is to build online systems with verified guarantees. Opportunities include guaranteed bounds on privacy, security, and performance.
Bio: Caleb Stanford is completing his PhD at the University of Pennsylvania (spring 2022), and holds an ScB from Brown University (2016). He has held industrial research internships at Microsoft Research Redmond in the Research in Software Engineering Group (2020) and Amazon Web Services in the Automated Reasoning Group (2019). Caleb's research focuses on programming languages and formal methods for systems applications, especially distributed systems, data processing systems, and automated verification tools. His work has been published at venues including PLDI, POPL, OOPSLA, and PPoPP. He also co-founded the Utah Math Olympiad, a high school proof-based math contest. At Penn, he holds a teaching certificate from the Center for Teaching and Learning, taught the Rust Programming course, and co-chaired the CIS doctoral student association for 4 years.
This seminar will be streaming via Zoom at the link above but requires a passcode. To obtain the passcode for this event, please see the announcements on the college email lists or contact us.