I'm Ilya. I build ML and AI systems — and the engineers who run them.
I've spent 15 years building AI/ML systems: leading ML for Adobe Analytics, owning model evaluations at Twitter, running a Staff-level team on Meta Ads, and managing Applied AI at Shopify — where I was also a hiring manager.
I recently spent a year back in the trenches inside a production AI org, so my advice is current — not a memory of how hiring worked in 2019. In the last six months alone I've made 50+ hiring decisions as a hiring-committee member.
Since 2020 I've helped 1,000+ engineers break in, clear interviews, and get promoted — through the newsletter, YouTube, and the programs.
The 5-step transition path
Your journey is your own, but I have seen these five steps work again and again. The trick is to do them in order.
Does an ML career actually fit you? What do you hope to achieve by moving into AI/ML?
Do a project you care about. Find the real challenges and frame your future learning.
Now take the class. Only now do you have enough context for it to be useful.
Show you can do ML in situations that matter — a validated project, not a toy.
Know which level and tier to target, then prepare. Preparation offsets less experience.
The 5 interview categories
Every ML loop is assembled from five kinds of rounds — and each has a different ideal preparation strategy.
Pattern recognition plus fluency in your language. Rewards consistent effort over cramming.
Best prep is designing systems. A few key concepts — and mocks — elevate your game.
Code up a model or fix a bug in one. Knowing a few libraries well goes a long way.
Any statistics or ML question is fair game. Understanding here makes you a better MLE, period.
Storytelling and listening — the HM round, the project deep dive. In other words: be human.