Why I Moved to Nebius

1. After ten years at AWS, Kubernetes work was starting to repeat itself. After writing a book on this subject, I wanted problems the book couldn't have covered yet.
2. Most of my business was already coming from customers trying to run serious AI and ML on Kubernetes.
3. I got tired of watching customer conversations get pulled toward Bedrock when many teams really wanted to run their own stack.
4. I wanted to work where AI infrastructure is the company, not one product line inside a giant cloud.
5. Less service noise means I can go deep. I spend more time on actual ML performance problems and less time trying to keep up with an endless catalog.
6. I wanted to write production code again. Not just architect it.
7. I missed real collaboration. At AWS too much of my time was either spent in meetings or alone on projects. Now I'm talking to teammates constantly, and it shows in the work.
8. The scale is different. Some of these customers are thinking in GPU fleets so large that ordinary cluster planning starts to look small.
9. Nebius is still a startup. That means the scope for influence is much bigger. AWS taught me how to build and operate at scale. I wanted a place where I could apply those lessons more directly.
10. I wanted to spend time with customers building what comes next not just modernizing what they already had.

I did not leave because I was done with AWS. I left because I wanted the next chapter to be at the part of tech where the answers are not settled yet. And right now, that means AI infrastructure, at a company where that is the mission.