Illia Polosukhin

From CryptoWiki


"My background is actually in machine learning. I worked for about 10 years at a few places. Specifically, my last big company gig was at Google Research, where I led a team working on question answering, machine translation. I was always really excited about open source, a lot of machine learning is open source and actually was one of the biggest contributors at the time to TensorFlow when it was just getting released by Google. For people who don't know, TensorFlow is probably like, top five open source projects on GitHub, in the world overall. 

So that was really exciting actually working in a big company, and at the same time really releasing something into the world, and really contributing and also working together with a lot of people around the world. There are really cool things where people would use TensorFlow for like classifying random things at their homes. 

But in general, Google, even though inside, it's a very open place, but it has a lot of constraints just being a big company. I left it to start a startup because I wanted to leverage things that I learned and some of the machine learning tools in the wider world. Me and my co-founder, Alex, we started the company called Near AI, which was supposed to teach machines to code. So we really wanted to have a machine learning model that a normal person who does not know how to do programming would be able to explain what they want, and the computer would write codes for them. It's a very challenging task, a lot of people have been working on it."