ML platforms for large companies: from DevBoxes to a platform for a thousand users
infra.conf 2026 (Yandex Infrastructure), Moscow
Public talks
22 talks: conference talks, podcasts, webinars and panels.
2026
Platformizing AI systems: problems of large and small companies
MLechny Put 2026 (Selectel ML meetup), Moscow
2025
Career growth in MLOps (with Yuri Klassen, Head of MLOps at Kuper)
ITMO x IT Infrastructure Club (podcast)
Expert: self x Zvuk career meetup, Moscow
self x Zvuk career meetup, Moscow
Choosing infrastructure for LLMs
Ya pro backend (Yandex), Moscow
GPU inference in K8s: acceleration, sharing and scaling without the pain
DevOops 2025 (JUG Ru Group)
NVIDIA Triton Inference Server: building production ML without developers
Recording for the Evrone YouTube channel (best talks playlist)
Inference autoscaling in K8s
ODS Data Fest 2025, Selectel day, Saint Petersburg
Taming LLMs: choosing inference infrastructure without the headache
MLechny Put 2025 (Selectel ML meetup), Saint Petersburg
Inference in ML: when you need NVIDIA Triton Inference Server and what else is out there (with Efim Golovin)
Data Breakfast (ODS) podcast, ep. 35
Inference autoscaling in Kubernetes
Ostrovok DevOps meetup (ProIT Fest), Moscow
Is the stage the key to success? How to succeed in IT (with Serafima Chekulaeva)
"Segodnya na retro" (Selectel podcast)
Efficient GPU utilization for inference and training (with Efim Golovin)
Data Breakfast (ODS) podcast, ep. 28 (announced as "episode three")
2024
NVIDIA Triton Inference Server: building production ML without developers
HighLoad++ 2024, Moscow
How to launch an inference platform with autoscaling, blackjack and live updates
Selectel Tech Day 2024, Moscow
Production ML: how to set up GPU nodes in Kubernetes clusters
Selectel webinar (online)
DevOps engineer (guest: Anton Alekseev, DevOps engineer for Data and ML products at Selectel)
"Rabotnik mesyatsa" (Red Barn podcast)
Panel: "DevOps vs. sysadmin: what's the difference and where to grow in 2024?"
Selectel Admin Meetup, Saint Petersburg
The unbearable lightness of raising GPU utilization: allocation and profiling for regular people
MLechny Put 2024 (Selectel ML meetup), Saint Petersburg
DevOps without ClickOps: how Terraform and GitLab made our infrastructure deploys 20x faster
DevOpsConf 2024 (Ontico), Moscow
2023
Speeding up with GitOps: automating workflows
Selectel Admin Meetup "Manage, simplify, automate!", Saint Petersburg