Today, we confront the persistent challenge of data sharing with privacy—a delicate balance where sharing valuable insights meets the imperative of safeguarding individual privacy. Differential Privacy offers a formal framework for reasoning about the privacy and accuracy of computations on private data. When carefully calibrated, these analyses guarantee the privacy of individuals and the accuracy of data analysis results, allowing inference of useful properties about a population. However, existing programming languages often fall short, providing support for reasoning about privacy but neglecting the critical aspect of accuracy in data analyses. To address this gap, we introduce DPella – an embedded domain-specific language (eDSL) in Haskell designed to empower data analysts with the ability to reason about privacy, accuracy, and their intricate trade-offs. A distinguishing feature of DPella is a novel component that statically tracks the accuracy of different data analyses using taint analysis, automatically inferring statistical independence to provide better error bounds. Following the birth of DPella, a startup bearing the same name emerged to bring the eDSL to the market. The second part of the talk will dive into some reflective thoughts and lessons learned by DPella (the company) when taking programming languages research into practice, from both the technical and the business aspects.

Information about the speaker: Alejandro Russo is a professor at Chalmers University of Technology / Göteborg University, working on the intersection of functional languages, security, privacy, and systems. His research ranges from foundational aspects of security to practical ones. Prof. Russo worked in prestigious research institutions like Stanford University, where he was appointed visiting associate professor back in 2013-2015. He has received the Most Influential Paper Award at ACM SIGPLAN ICFP 2022 and was recognized with the Distinguished Paper Award at ACM SIGPLAN POPL 2019. In 2020, he co-founded DPella, a startup providing Differential Privacy solutions in Haskell, where he is currently the CEO. In addition to his academic pursuits, Alejandro is currently a scholar in an Executive MBA program (2022-2024).