Cedric Renggli

I am a Senior Researcher and Lecturer at ETH’s Systems Group, working with Ana Klimovic.

My research lies at the intersection of data management and AI/ML systems. I mainly study the data that AI systems consume and produce, and how to manage it so that these systems require less compute and data while the quality of their results stays measurable, ideally with statistical guarantees. During my PhD at ETH Zurich, I focused on the quality of static datasets, through feasibility studies based on Bayes error estimation and continuous integration of ML models with statistical guarantees at minimal labelling cost. At Apple, I turned to operational data, leading the design of on-device vector search and studying LLM-based access to relational data (Text2SQL). Today, I focus on the interaction traces that AI agents generate at runtime: managing them as a new class of data, and using them to build workload-aware serving systems, estimate usage with statistical guarantees, and control cost and quality in unreliable execution environments.

Previously, I was a Senior Researcher at Apple and a PostDoc at UZH (DaST, with Dan Olteanu). I received my PhD from ETH’s Systems Group, advised by Ce Zhang.