Giulia Fanti

Timeline

1. 2010–2015

   Olin and Berkeley degrees

   She earns a BS in electrical and computer engineering at Olin College of Engineering in 2010. In 2015, she completes a PhD in electrical engineering and computer sciences at the University of California, Berkeley, advised by Kannan Ramchandran; her thesis examines anonymous message spreading and private database searches.

2. 2017

   Joining Carnegie Mellon

   Carnegie Mellon welcomes her to its Department of Electrical and Computer Engineering faculty in 2017, while she finishes a postdoctoral appointment at the University of Illinois at Urbana-Champaign. Originally from Lexington, Kentucky, she studies privacy and anonymity.

3. 2018

   PacGAN and sample diversity

   With Zinan Lin, Ashish Khetan, and Sewoong Oh, she coauthors PacGAN at NeurIPS in 2018. The method gives a generative adversarial network's discriminator several samples at once to reduce mode collapse, a failure in which generated data lack variety.

4. 2020

   DoppelGANger for synthetic data

   In 2020, she coauthors the DoppelGANger study presented at the ACM Internet Measurement Conference, with researchers at Carnegie Mellon's CyLab and IBM. The tool uses generative adversarial networks to make synthetic datasets with the same statistics as training data, supporting data sharing and training for network security engineers.

5. 2021

   A Sloan Research Fellowship

   She receives a Sloan Research Fellowship in 2021. Her research examines the security and privacy implications of data transparency and sharing, spanning distributed systems, machine learning, and privacy technologies.

6. June 2024

   Co-founder of Rockfish Data

   By June 2024, she is co-founder and chief scientist of Rockfish Data, one of the startups in Carnegie Mellon's CyLab Venture Network.

7. July 2024

   The Angel Jordan professorship

   In July 2024, she becomes the Angel Jordan Associate Professor of Electrical and Computer Engineering at Carnegie Mellon. She is also a founding co-director of CyLab-Africa, which focuses on the security and resilience of Africa's financial technologies and infrastructure.

8. 2025

   POPri for private learning

   In 2025, she coauthors Private Federated Learning using Preference-Optimized Synthetic Data at ICML with Charlie Hou, Mei-Yu Wang, Yige Zhu, and Daniel Lazar. Their POPri algorithm uses client feedback to fine-tune language models that generate synthetic data with differential privacy.

9. As of October 2026

   Professor and Upanzi co-director

   In March 2026, she visits the Simons Institute at the University of California, Berkeley, for its Federated and Collaborative Learning program. As of October 2026, she is the Angel Jordan Associate Professor at Carnegie Mellon, with courtesy appointments in Computer Science and Machine Learning and membership in CyLab. She also co-directs the Upanzi Network at CMU-Africa and the Initiative for Cryptocurrencies and Contracts, and is a senior fellow at the Atlantic Council GeoEconomics Center.