Giuseppe Vietri

I am an applied research scientist, based in New York City. I work on synthetic data generation for LLM post-training and evaluation using novelty search — an approach that optimizes for diversity under task-specific constraints rather than fixing a single reward. My work spans private tabular synthesis, prompt augmentation, and teacher-student distillation. Ph.D. in differential privacy for machine learning (University of Minnesota, 2023).

I obtained my Ph.D. in 2023 at the University of Minnesota, Twin-cities in the Computer Science and Engineering Department. I was co-advised by Maria Gini and Steven Wu. Before that, I completed a Master's in Computer Science at Florida International University under Professor Giri Narasimhan.