Build log.
Notes on full-stack development, backend engineering, and building real things with AI — written as I learn them, not after the fact.
Why I Still Read Every Line an AI Writes for Me
AI can write code faster than I ever could, but that doesn't mean I should stop understanding it. In an age where generating software is becoming easier, engineering judgment matters more than ever. Here's why I still read every line AI writes for me—and why I believe developers should never outsource their understanding.
A Software Developer Should Think in Systems, Not Screens
Modern software development makes it easy to focus on screens, components, and features—but the real engineering happens beneath the interface. In this article, I explore why developers should think in systems: understanding data flows, dependencies, failure modes, feedback loops, scalability, and the long-term behavior of the software they build. Because a screen is only the surface; the system behind it is what makes software reliable.
Lessons from Building My First RAG Pipeline
Building my first RAG pipeline taught me that successful AI applications depend on much more than choosing a powerful LLM. From document chunking and embeddings to retrieval quality and evaluation, every stage affects the final answer. These are the practical lessons I learned while turning a simple RAG prototype into a more reliable AI system.