AI-generated deep dives into software engineering — from frontend architecture and best practices to technical leadership and team building. Each episode explores real-world lessons from building large-scale applications. Audio generated using Google NotebookLM, topics curated from hands-on engineering experience.
Reading the feed…
Redux, state machines, saga middleware — most teams don't need any of it. React state and Context solve 90% of real-world cases, and the remaining 10% probably don't need a library either.
Most React developers write code that works but re-renders 10x more than it should. Understanding Fiber, reconciliation, and when useMemo actually matters changes how you think about performance.
How I built a conversational AI agent with Claude Haiku, a single Next.js API route, and zero npm dependencies. Rate limiting, markdown rendering, syntax highlighting, and privacy guardrails — from idea to production in under an hour.
A real-world multi-step form with validation, i18n, and event-driven architecture — built with Lit, native HTML validation, and 5 dependencies. No React. No state library. No form library.
If your small Scrum team runs a Java Spring Boot BFF alongside a React frontend, you're paying double. Here's an honest architecture comparison — including where Java wins.
The W3C told us in 2006: use the least powerful technology that works. Most frontend teams do the opposite. Here's how HTML, CSS, and native APIs can replace half your dependencies.
"Because it's a best practice" is not an argument — it's an appeal to authority. Why following blindly is dangerous and how to build your own practices.
The most impactful architectural decision in frontend: separating what fetches data from what renders it.
A hands-on look at using AI as a development partner — from Figma design to production code with Claude.
Most developers learn DRY early and apply it everywhere. But premature abstraction kills more codebases than duplication ever will.
What actually works when leading frontend teams at scale — from code review as mentorship to hiring engineers who think, not follow.
Reduce your Node dependencies for maintainable apps. Every package you add is a liability — learn when to use them and when to write it yourself.
I designed a micro-frontend architecture at AXA Switzerland. Here's what I learned about when this pattern helps and when it's just complexity for complexity's sake.
