Gerome Miklau
Adjunct Professor
Manning College of Information and Computer Sciences
University of Massachusetts, Amherst
Principal Staff Software Engineer
LinkedIn Corporation
I am an Adjunct Professor of Computer Science at the University of Massachusetts, Amherst, and a Principal Staff Software Engineer at LinkedIn. I was previously a full-time Professor (2005–2022) and co-founded Tumult Labs, a start-up focused on commercializing privacy technology. My research advances privacy-preserving AI and the fair, responsible use of data, through differentially private algorithms and systems, synthetic data generation, and work on algorithmic fairness and transparency.
Full Biography
Gerome Miklau is an Adjunct Professor of Computer Science at the University of Massachusetts, Amherst, and a Principal Staff Software Engineer at LinkedIn. Previously a Professor at UMass, his research advances privacy-preserving AI and the fair, responsible use of data. He designs algorithms and open-source systems for differentially private data analysis, sharing and release, including high-performing synthetic-data generation methods.
He co-founded Tumult Labs in 2019, a start-up focused on commercializing privacy technology, and served as its CEO from 2022 through its acquisition by LinkedIn in 2025. Prior to that, he consulted for the U.S. Census Bureau on algorithms deployed for the 2020 decennial census.
Miklau received the ACM PODS Alberto O. Mendelzon Test-of-Time Award in both 2020 and 2012, the Best Paper Award at the International Conference of Database Theory (ICDT) in 2013 and the ICDT Test-of-Time Award in 2023, a Lilly Teaching Fellowship in 2011, an NSF CAREER Award in 2007, and the 2006 ACM SIGMOD Dissertation Award. He received his Ph.D. in Computer Science from the University of Washington in 2005. He earned Bachelor's degrees in Mathematics and in Rhetoric from the University of California, Berkeley, in 1995.
Research Highlights
Open-source systems for private computation
I build open-source systems that make differentially private computation practical. This includes PrivateSQL, a differentially private SQL query engine, and Ektelo, a system for defining differentially private computations.
Tumult Labs developed Tumult Core and Tumult Analytics, building on this work. These open-source privacy frameworks were deployed by U.S. government agencies and private-sector organizations, then donated to OpenDP, where they remain in use and are actively maintained.
Synthetic data generation
Our team developed award-winning algorithms for generating accurate structured synthetic data with provable privacy, including multiple prize-winning entries in the NIST Differential Privacy Synthetic Data Challenge. These methods use probabilistic graphical-model estimation and inference as a key subroutine; AIM remains one of the leading approaches to differentially private synthetic data generation.
Accurate private query answering
The Matrix Mechanism is one influential approach from my work on improving the accuracy of statistical queries under differential privacy. Introduced in our 2010 PODS paper, it strategically selects which queries to measure. The paper received the 2020 ACM PODS Alberto O. Mendelzon Test-of-Time Award; subsequent work extended the approach to high-dimensional statistical queries and produced a public HDMM implementation.
Algorithmic fairness
I study fairness and transparency in data-driven decision-making, including fair decisions using privacy-protected data and tools for transparent, stable, and responsible ranking design.
Applications of private data analysis
My work applies privacy-preserving data analysis to diverse domains, including rank aggregation, data markets and the pricing of private data, private network analysis, and visual analysis of differentially private data. I have also developed methods for benchmarking privacy algorithms and integrating public information into private data analysis.
Students
I recently served on the PhD committees of Brett Mullins (Allegheny College), Arisa Tajima (LinkedIn), Cecilia Ferrando (LinkedIn), and Miguel Fuentes (Capital One), who have since graduated.
My former PhD students are Ryan McKenna (2022, Google Research), Dan Zhang (2021, Megagon), Yue Wang (2017, Microsoft Research), Wentian Lu (2014, Google), Chao Li (2013, Google), and Michael Hay (2010, Colgate University).
Selected Service
I currently serve as co-chair of the OpenDP Advisory Board, an open-source project for differential privacy to which Tumult Labs donated its core technology.
I previously served on the steering committees for the Theory and Practice of Differential Privacy (TPDP) workshop series and the Conference on Fairness, Accountability, and Transparency (FAccT). I co-organized the "Data, Responsibly" workshop at Schloss Dagstuhl and served as Associate Editor for Transactions on Knowledge and Data Engineering (TKDE).
I have served on the program committees of a variety of computer science conferences and workshops including SIGMOD, PODS, PVLDB, ICML, TPDP, FAccT, ICDE, ICDM, CCS, WebDB, and CIKM.