Beyond Vibe Coding: Development Productivity Without the Slop Code
Learn how to orchestrate AI with the Cognitive Rebase method instead of manually debugging generated code. Focus on the big picture and the interesting problems while confidently delegating the details to AI.
2 hours each, online, live
Hands-on, project-based learning
For teams building new software
What You'll Learn
A practical workshop — 10% theory, 90% practice. The core goal is building a personal finance tracking system from scratch with AI, using TypeScript / Node.js / React. Over 4 sessions of 2 hours each, you'll work on your own project — task by task — following three core approaches:
Test-Driven Development
Express your requirements as executable specs — tests that give clear criteria for whether AI is generating code in the right direction or not.
Small Steps
Limit the blast radius of hallucinations. The smaller the steps, the more control you have over AI.
Modular Architecture
Follow Hexagonal Architecture and Domain-Driven Design to constrain AI into a business framework, limiting changes to what matters for the current task.
You'll also explore spec-driven development (Tessl, Kiro), ground rules (CLAUDE.md), MCP servers for context from external tools (Figma, GitHub, Jira), and other tools in the ecosystem (CodeRabbit, CommandCenter, Devin, Factory, Lovable, Replit).
Session Breakdown
Executable Specifications
How executable specs help us think about the important problems
- - How to structure your project for easy maintenance of executable specs
- - Implementing first features following tight TDD feedback loops
- - The value of executable specs for catching regressions
- - Challenge: even with executable specs, AI often goes down rabbit holes
OUTCOME: You'll learn how to direct AI to create exactly what you want through tests that the business understands and are easy to maintain.
Small Steps & Executable Specs
How working in small steps combined with executable specs helps navigate AI better
- - Creating a text specification and breaking it into vertical slices
- - Selecting a slice and approaching it in miniature steps
- - The role of refactoring after every miniature step
- - Challenge: refactoring solves local design problems
OUTCOME: You'll learn how to prevent AI from accumulating errors and the time cost they bring.
Modular Architecture
How modular architecture keeps you focused on the big picture while delegating details to AI
- - What Hexagonal Architecture and Domain-Driven Design look like in practice
- - Laying the architectural foundations for the personal finance app
- - Navigating AI within architectural boundaries and full delegation of implementation
OUTCOME: You'll learn how focusing on business logic (the big picture) helps you delegate implementation details to AI.
Tooling & Ecosystem
Additional tools that will make you more productive
- - Spec-driven development tools: Kiro, Tessl, GitHub Spec-kit
- - Integration with GitHub MCP server
- - Easy code review with CodeRabbit and CommandCenter
- - Working at scale with Factory and Devin
- - Rapid prototyping with Lovable and Replit
OUTCOME: You'll learn which tool is missing from your arsenal to accelerate your software development process.
Key Takeaways
- An engineering approach to using AI in the software development process
- How to avoid hallucinations and control AI to do what you want
- How to stay focused on the big picture while delegating details to AI
- Practices and tools that make you an AI-Native Software Engineer
- A project you built entirely with AI as proof of concept
Ready to transform how your engineering team works with AI?
Book a 30-Minute Fit Call