Illia Polosukhin
Illia Polosukhin is a Ukrainian computer scientist known as the eighth and final co-author on Attention Is All You Need, the 2017 paper that introduced the Transformer architecture, and as a co-founder of NEAR Protocol. He occupies a rare niche in the technology world: someone who helped light the fuse on the modern Large Language Model boom and then walked straight into blockchain to build something new with it.
From Kharkiv to Google Brain
Polosukhin grew up in Kharkiv, Ukraine, and studied applied mathematics before joining Google in 2014 as a software engineer on the search team. At Google, he worked on question-answering systems and contributed to TensorFlow, Google's open-source machine learning framework, while immersing himself in the currents of Neural Machine Translation and Sequence To Sequence Learning that were reshaping natural language processing.
Attention Is All You Need
In 2017, alongside Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan Gomez, and Łukasz Kaiser, Polosukhin published the paper that dispensed with Recurrence and Convolutions in favor of pure Self-Attention. His contribution centered on the machine-translation experiments that first demonstrated the architecture's value before it was generalized to other language tasks. The resulting Transformer became the scaffolding beneath BERT, GPT, GPT-3, T5, and the Vision Transformer — an entire family tree of models built on a single insight about Attention.
Before Illia Polosukhin left Google in 2017, he had a brainstorming lunch and then returned to his desk to build what may have been the very first transformer, a moment now retold as a small origin story within the larger mythology of the AI And LLM Papers Reading Map.
Building NEAR Protocol
Polosukhin co-founded NEAR Protocol with Alexander Skidanov in 2018, later serving as its CEO. NEAR is a Layer 1 blockchain designed for usability and scalability, featuring a sharded architecture called Nightshade that allows the network to scale throughput by splitting processing across multiple shards. He has since become a leading voice on the convergence of AI and blockchain technology, advocating for decentralized AI governance and on-chain AI agents.