← Journal · · Announcement
We wrote a course on MLOps!
I want to share a project I've put a lot of effort into over the past year: the «MLOps for Model Development and Monitoring» course at Yandex Practicum. 🚀🚀🚀
For me, this wasn't just a "help out with the course" gig - I put together the program core to match the current market (and my own experience), wrote specs for the authors, and designed the course infrastructure. It was hundreds of hours of work: rebuilding the program, refactoring modules, deadlines, approvals - basically real product development, not just a stack of lectures.
And personally, this experience gave me a lot too - networking and meeting strong authors (thanks to everyone for their contribution, and especially to Katya Tsaplina for the dev and ML practicals in the course, and for inviting me onto the project!), the experience of building a large educational product, a chance to pack my practical view of MLOps into a proper learning path, and, honestly, just a nice feeling that a year ago I promised to write a course and actually ended up doing it 😅
What matters about the course itself: I wanted to make it not just "another MLOps program from data scientists," but a course with a strong focus on the infrastructure side. There are plenty of courses on the market where MLOps is presented mostly from the Data Science angle, using ready-made tools. I wanted to show what it looks like from the perspective of an engineer who builds and operates the platform.
So what will a student end up with:
- an understanding of the full MLOps cycle (see one and two)
- skills for the Dev, ML and OPS tracks of an MLOps specialist (according to the diagram)
- hands-on practice with ClearML, Kubernetes, MLflow, Airflow
- inference on K8s and on KServe
- observability of models and infrastructure
- data quality and pipeline quality control
So the idea of the course is that afterward a developer / DevOps engineer / data scientist can move much more confidently toward junior MLOps and understand not just the models, but the whole production loop around them. One thing I'll flag upfront: the course doesn't cover GPU work. It just didn't fit, and I think that's the next level anyway (or more relevant to LLMOps).
There's also a personal side to this for me: once upon a time I took Practicum's own DevOps course, and it had a big influence on my track back then, it helped me find the job where I became an MLOps engineer. So it's especially cool that now I ended up on the other side and built a course of my own.
If you want to sign up, message me directly - I'll share a 15% promo code.
And for those who buy the course with this promo code, I'm happy to gift a half-hour mentoring session with me on getMentor: we'll talk about a career track in MLOps, what to focus on first and how to get in - leave a request with a note that it's from the MLOps course ❤️