Articles
Insights from the frontier.
Practical AI, product delivery, and engineering execution.
17 articles
May 18, 20263 min read
Don't pre-build the AI layer
The retail industry has a name for the developer who builds the mall before the anchor tenant signs. Optimistic. The same developer five years later, when the mall is half empty and the anchor never showed, is called bankrupt.
May 8, 20263 min read
We added the meeting
Once, somebody had the bright idea that the team was not close enough. Let’s call this idea `the Standup`. Any resemblance to actual meetings is purely coincidental. Probably
May 6, 20263 min read
The billable hour was always a workaround
In the early 1900s, coal miners across South Wales were paid by the ton. It was a reasonable arrangement. Effort and output correlated tightly enough, measurement was simple, and the variance between a fast miner and a slow one was narrow enough that the proxy held.
May 5, 20263 min read
Not everything needs an agent
In 1931, Rube Goldberg won a Pulitzer Prize for drawings of machines that accomplished simple tasks through spectacular chains of unnecessary steps. A self-wiping napkin apparatus involving a parrot, a lit candle, a swinging pendulum, and seven other components.
May 5, 20263 min read
The million-token search engine that isn’t
A library with no card catalog has every book you could want and no way to find any of them. Bigger libraries don’t fix search. They make search worse. That is most of what is wrong with the current marketing of long context windows.
May 5, 20263 min read
Schema before prompt
A stage director can rehearse the play to perfection. The actors hit their marks. The lighting cues are timed to the breath. But if the prop master walks out with a sword when the script calls for a letter, the scene collapses, and the audience walks out blaming the actor.`
May 5, 20263 min read
They skipped the boring part
Last year, we hired a junior engineer with fifteen GitHub repositories, three of them with actual stars. He shipped fast, had opinions about architecture on day one, and could produce working code in the time it took me to finish my coffee. We were delighted.
Jan 8, 20267 min read
Perfect Way of Versioning Models & Training Data
Struggling with versioning training data and ML models? This article introduces DVC (Data Version Control) and shows how to properly version datasets and model files alongside Git. Learn hands-on workflows, real-world use cases, and how to collaborate efficiently in MLOps projects.
Jan 8, 20266 min read
Vibe Coding: The Web Development Approach You Didn’t Know You Needed
Discover Vibe Coding—a fast, idea-first approach to web development powered by AI tools like v0. Learn how to generate production-ready React apps in minutes, reduce boilerplate, skip overengineering, and bring the joy back to building modern web interfaces.
Jan 8, 20268 min read
Next.js 15 Whats new?
Explore the Next.js 15 Release Candidate, featuring React 19 RC support, major caching behavior updates, Partial Prerendering, the new after() API, and experimental React Compiler integration. Learn how these changes improve performance, rendering control, and serverless workflows.
Jan 8, 20263 min read
Mimicking React’s useState Hook with JavaScript Proxy Object
Learn how to recreate React’s useState-style reactivity in vanilla JavaScript using the Proxy object. This hands-on guide shows how to auto-update the UI, manage state changes, and add validations—without any frameworks.
Jan 8, 20263 min read
Building “TagTamer” — A Chrome Extension with Manifest V3
Learn how to build a Chrome extension using Manifest V3 by creating TagTamer—a simple tool that highlights HTML heading tags on any webpage. This hands-on guide covers popup UI, background scripts, content scripts, and modern extension architecture.
Jan 8, 202610 min read
Build Interactive Data Dashboards and Talk to Data Using Vizro & Vizro-AI
Learn how to build interactive Python dashboards with Vizro and use natural language to explore data using Vizro-AI. This step-by-step guide covers EDA, modular dashboards, Plotly visualizations, and AI-powered data analysis for modern data science workflows.
Jan 7, 20268 min read
Accelerate Your Machine Learning: A Hands-On Introduction to RAPIDS in Pyth
Struggling with slow pandas and CPU-bound ML workflows? This hands-on guide introduces RAPIDS, showing how GPU-accelerated libraries like cuDF and cuML can drastically speed up data processing, model training, and large-scale ML tasks in Python.
Jan 7, 20265 min read
Taking your RAG pipelines to a next level ! LangGraphs
Already working with RAG pipelines and LangChain? This article introduces LangGraph, LangChain’s latest upgrade, and shows how it elevates RAG workflows with better structure, control, and scalability for modern generative AI applications.
Jan 7, 20268 min read
When Data Lies: Exploring the Enigmatic World of Data Science Paradoxes
Data doesn’t always tell the truth at first glance. This article explores famous data science paradoxes—like Simpson’s, Berkson’s, and the Accuracy Paradox—to help you uncover hidden variables, avoid misleading conclusions, and think more critically about real-world data.
Jan 7, 202610 min read
Stop Writing Complex Code: Build Smarter Workflows with AWS Step Functions
Building AI and data pipelines doesn’t have to mean tangled code and fragile workflows. In this article, we explore how AWS Step Functions can simplify orchestration, improve reliability, and help you build scalable, serverless workflows, without the headache of managing complex logic yourself.
Newsletter
Stay in the loop.
New essays and field notes on AI delivery, engineering leadership, and product execution — straight to your inbox.