Super Short Bio

Hi, I’m Pavel. I am currently exploring the intersection of applied AI, tools for scientific exploration, and industry. I live in Berlin, Germany, with my wife and daughter.

More about me

Selected Peer-reviewed Publications

• E. Lagutin, D. Gavrilov, P. Kalaydin, Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning¹, Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics, 2021.

• N. Zueva, M. Kabirova, P. Kalaydin, Reducing unintended identity bias in Russian hate speech detection, Proceedings of the Fourth Workshop on Online Abuse and Harms, 2020.

• D. Gavrilov, P. Kalaydin and V. Malykh, Self-Attentive Model for Headline Generation, Proceedings of the 41st European Conference on Information Retrieval, 2019.

Recent Interviews/Opinions

• Constructor University und Firma Sumsub forschen zu Deepfakes: Weser Kurier (22.09.2025)

• TechTransformed: The AI Fraud Epidemic: Insights from Global Trends, an episode on AI Fraud: Apple Podcasts (05.08.2024)

• My commentary on deepfakes and identity fraud in Bloomberg (18.06.2024)

• African elections in deepfakes, live, at CNBC Africa (31.05.2024)

Selected Talks

• Money2020: AI vs. Digital Fraud: Unmasking Fintech’s Hidden Enemy, Money2020, Bangkok, 2024

• AI as a Catalyst for Disruption and Innovation Across Industries, panel discussion, GTC 2022

• Deploying Natural Language Processing Beyond English, panel discussion, GTC 2021

• The Allen AI Science Challenge (7th place) & DeepHack Q&A hackathon (1st place) (slides)

Selected Side Projects

• LLMArena, a text game where you prompt-inject a model to win in a combat (itch.io profile)²whuno