Founder & Member of Technical Staff at DatologyAI
Pratyush Maini is a researcher with a PhD in Machine Learning from Carnegie Mellon University who works on training data, memorization, and safety of foundation models.
How media typically covers Pratyush Maini
Pratyush Maini as author
BeyondWeb, a synthetic data generation framework, outperforms state-of-the-art synthetic pretraining datasets like Cosmopedia and Nemotron-Synth by up to 5.1 percentage points while enabling 7.7x faster training than open web data.
“Author of "The BeyondWeb Paper" on arXiv”