Illia Polosukhin
TL;DR AI Pioneer and Blockchain Visionary shaping the future of decentralized systems, co-author of “Attention Is All You Need” and co-founder of NEAR Foundation.
Illia Polosukhin by Sora
Illia Polosukhin is a Ukrainian AI researcher and software engineer who has helped shape modern artificial intelligence and blockchain technology. He was part of the team that published a landmark 2017 paper that introduced the transformer architecture, the foundation of generative AI. After leaving Google, he co‑founded the NEAR project and now leads the NEAR Foundation, using his expertise to build scalable, open infrastructure for decentralized applications.
Born and educated in Ukraine, Polosukhin began his career developing tools for big‑data analytics and text mining. At Google Research, he worked on TensorFlow and led teams advancing deep learning and natural language understanding. The transformer architecture he co‑developed uses self‑attention to process data efficiently and has become the backbone of modern large language models.
In 2017, he and Alexander Skidanov launched NEAR.ai to explore AI and program synthesis before pivoting to blockchain. Inspired by principles from database systems and language model training, they designed NEAR Protocol, a layer‑1 blockchain with sharding to solve the scalability trilemma. NEAR launched in 2020 and has raised over $550 million, enabling user‑friendly, scalable decentralized applications.
Polosukhin advocates a vision of user‑owned, transparent AI rather than systems controlled by a few corporations. In March 2025, he presented research at NVIDIA’s GTC conference on decentralized confidential machine learning that preserves user privacy while enabling developers to monetize models. Appointed CEO of the NEAR Foundation in November 2023, Polosukhin now guides the ecosystem toward an open web.
Co‑authored the 2017 paper that introduced the transformer architecture and sparked modern generative AI.
Led deep learning and natural language processing teams at Google and contributed to TensorFlow.
Co‑founded NEAR.ai and pivoted it into NEAR Protocol, a scalable layer‑1 blockchain.
Raised over $550 million in funding for NEAR Protocol.
Championed the concept of user‑owned AI and open, transparent models.
Presented research on decentralized confidential machine learning at NVIDIA GTC 2025.
Appointed CEO of the NEAR Foundation to guide the platform’s evolution toward an open web.
Featured by major media outlets such as Wired, CNBC, and TIME.
