Dr. Artem Shelmanov / Assistant Professor of Practice

I am an AI researcher committed to conducting high-impact research in natural language processing and artificial intelligence that drives positive change in the world.

At Mohamed bin Zayed University of Artificial Intelligence: MBZUAI (ranked 10th in AI globally), I work in the research group of Prof. Timothy Baldwin on one of the critical challenges in AI – LLM reliability. As a team leader, I drive research into robust uncertainty quantification (UQ) methods, aiming to effectively detect and mitigate LLM hallucinations. Our research aims to increase the applicability of AI technologies in safety-critical areas, such as healthcare and finance, as well as increase the performance of multi-step reasoning systems such as LLM agents.

My team has created the most comprehensive UQ library for LLMs: LM-Polygraph.

We have presented our tutorial “Uncertainty Quantification for LLMs” at ACL-2025 in Vienna, AAAI-2026, and ECIR-2026. Our new tutorial, “Uncertainty Quantification: From Detecting LLM Hallucinations to Strengthening Reasoning and AI Agents,” has been accepted at NeurIPS-2026 in Sydney. It focuses on uncertainty as a control signal for reliable reasoning, adaptive test-time compute, and safer tool use.

Recent News

Graduated Students

  • Ekaterina Fadeeva — MSc student, 2023-2024 @ HSE University (co-supervised with Maxim Panov). Best Thesis Award on Uncertainty Quantification in LLMs. Now PhD student @ ETH Zurich with Prof. Mrinmaya Sachan.
  • Artem Vazhentsev — MSc & PhD student, 2021-2025 (co-supervised with Alexander Panchenko). Thesis on "Uncertainty Quantification in LLMs" defended at HSE University received cum laude distinction.
  • Gleb Kuzmin — PhD, 2026 @ HSE University (co-supervised with Prof. Ivan Smirnov). Successfully defended his thesis, "Uncertainty Estimation and Fairness in Text Analysis Tasks", in June 2026.