← Journal · · Personal
2025 wrap-up
Lots of people are actively wrapping up the year, so I decided not to stay on the sidelines and recall what happened with us this year :)
I split it into a few tracks:
🎤 Talks and appearances
- Learned together with you about inference scaling in K8s, GPU workload schedulers, and GPU sharing at ODS DataFest and DevOops Conf.
- Became true Vikings and tamed the LLM at MLechny Put and Yandex's conference - Ya pro Backend! At the latter I broke my own record - 200+ listeners and 56+ talk ratings. Thanks to every one of you ❤️
- Published 2 articles on Habr - about prepping Triton recipes and picking infrastructure for LLMs. Part 2 coming!
🧑🎓 Mentoring and MLOps growth
- Helped 13 mentees on the getMentor platform! The main topics people came to me with - MLOps roadmap, building platforms for ML workloads, GPU in K8s. Listen to the podcast about the MLOps track!
- Put together the curriculum for an MLOps course on a well-known platform. Since we're currently refactoring the first modules, I'll share the details next year)
📝 The channel
- A rebrand once a year, like clockwork. Curious what's next...
- 761 subscribers by year end! +450 subscribers over 2025, you're amazing!
- Great engagement rate on views! A full 97.5% - meaning on average every one of my subscribers reads every post!
- The legendary post unboxing an Nvidia B200 at 500 subscribers. Hoping to work with real ones next year.
👨 Personal
- Tried moving to Moscow, lived there for six months, and moved back to St. Petersburg. The movie Shrek opened up for me in a whole new way))
- Changed my main job over the summer. Now I develop the ML and inference platform at Avito 👌🆗
- Added a few new countries to my travel list - China, UAE, and Indonesia!
- Hit 1000 days on Duolingo 😂. Good thing the only abuse in my life comes from that bird.
Dear friends, really glad you keep reading this channel (even though posts have gotten rarer), support me, and speak warmly not just about the content but about what I do overall! Next year I promise not to slow down and to keep you entertained with posts about AI/ML infrastructure!
Happy New Year and Christmas to everyone ❤️