Shreyansh Padarha profile pic

Shreyansh Padarha

2026 Cohort

Trustworthy AI Scientists: Measuring and Safeguarding Epistemic Integrity in Autonomous Research

Shreyansh is a DPhil student at the University of Oxford, working on building methods to evaluate and improve how large language models take part in scientific research. He is interested in how far current models can support autonomous research, and how much human oversight is still needed to keep this safe and reliable. Alongside his primary research, he contributes to related projects on post-training methods, agentic safety and multimodal model evaluation. He is supervised by Professor Adam Mahdi and is part of the Reasoning with Machines Lab (OxRML).

He has long been interested in applied AI research, having worked across healthcare, agriculture and media. Before starting his DPhil, he was a research assistant at the Oxford Internet Institute, where alongside his own research he contributed to a EPSRC-funded project using LLM-based methods to identify and measure migration narratives in traditional and social media. He holds an MSc in Social Data Science (Distinction) from the University of Oxford, and a BSc in Data Science from Christ University, India (Gold Medallist).