← Caseseducation

Yandex Practicumsince 2025

The "MLOps for model development and monitoring" course

A course that shows MLOps through the eyes of an engineer who builds and runs the platform. I am the program expert for every module and was responsible for the curriculum.

Numbers

10modules, program expert for all of them
5+authors on the team
April 2026the course went live

Before

  • Many courses teach MLOps from the data science side and around ready-made tools.
  • A course that shows MLOps through the eyes of a platform engineer was missing.
  • Back in my 2024 year-end post I promised to launch my own course.

After

  • The course has been in the Practicum catalog since April 2026: 10 modules, the full MLOps cycle, Dev, ML and Ops tracks.
  • Graduates reach the junior MLOps level.
  • Hundreds of hours of work: rebuilding the program, refactoring modules, deadlines. Real product development, not a set of lectures.

What I did

  1. Built the core of the program for today's market down to modules and lessons.

  2. Wrote the briefs for the authors and led a team of 5+ authors, reviewed the materials.

  3. Designed the course infrastructure together with the DevOps team and optimized cloud costs per student.

  4. Focus on infrastructure: hands-on work with ClearML, Kubernetes, MLflow and Airflow, inference on Kubernetes and KServe, observability of models and infrastructure, data quality.

  5. GPUs were deliberately left out: that is the next level, closer to LLMOps.

What the course teaches

ClearMLKubernetesMLflowAirflowKServeobservabilitydata quality

My role

Program expert for every module of the course: responsible for the curriculum. Working on the course since July 2025.

Course page (in Russian) ↗

Discuss a similar taskTelegrama consultation via getMentor or Telegram

↑↓ select · Enter open · Esc close