Sergey Ris's Vibe Coding Education

Table of content

Sergey Ris is an engineer, designer, and educator who spent 10+ years in development and 8+ years teaching at Yandex Practicum. He co-founded HSL, a digital creative studio, and launched Vibe Coding School to bring vibe coding to Russian-speaking non-programmers. His Telegram community Vayb Codery has grown to 1,100+ members.

Background

Former instructor at Yandex Practicum, trained 500+ students
Design and development work at Yandex and Strelka architecture institute
Generative artist creating AI-driven installations and art

Find him on Telegram (@serejaris), Twitter (@riiiiiiiiss), or his YouTube channel.

The Framework

Ris teaches a four-phase workflow similar to the three-layer approach and Harper Reed’s methodology, but optimized for beginners who have never written code. See Claude Code for Non-Developers for context on this audience.

Phase	What Happens
Research	Understand the problem space before touching tools
Planning	Design the solution structure with AI assistance
Execution	Build in small batches, validate each step
Learning	Document what worked for the next project

The key difference from typical vibe coding tutorials: Ris emphasizes human agency at every step.

“The first step is made by a person. Not the model. Not the IDE. You.”

This framing matters for his audience. Non-programmers often defer entirely to AI output. Ris teaches them to stay in the driver’s seat.

The Vibe Coding School

The 4-week intensive covers the full AI-assisted development stack:

Tool	Use Case
Claude Code	Terminal-based development, full codebase access
Cursor	IDE integration, inline edits
Lovable	Visual prototyping
Bolt	Quick web app scaffolding

Students build real products, not toy examples. The curriculum assumes zero programming background but expects participants to ship something functional by week 4.

Philosophy

Ris sees a change in who can build software:

“For 60 years we taught machines. Now the machine learned our language.”

His approach centers on decisiveness over tools. The specific AI tool matters less than knowing what you want to build and committing to shipping it.

From his Telegram blog:

“Everyone with Claude Code is now a researcher and explorer.”

People who couldn’t build software a year ago can now ship production applications. That’s what he teaches.

Key Takeaways

Principle	Implementation
Human agency first	You decide direction, AI executes
Research before building	Understand the problem before prompting
Small batches	Validate each step before continuing
Ship real products	Toy examples don’t teach shipping
Document what works	Build a personal knowledge base (learning in public)

Links

Next: Andrej Karpathy’s AI-First Development