K-Means is a classic clustering algorithm in machine learning, and one of my favorite ways to teach it is through a hands-on project where students have to build an application that extracts a color palette from an image. The idea is simple, each pixel’s RGB values are treated as data points, and K-Means clusters them to identify the dominant colors. It’s a visual and intuitive way for students to understand how clustering works, and it typically takes a few hours to implement from scratch.

This weekend, while updating some curriculum and experimenting with coding agents, I was genuinely surprised by how quickly this same project came together. With the right prompts, the agent was able to generate a working version in just minutes. I also asked Nano Banana to generate a sample image to test the app, making the entire workflow from data generation to deployment extremely fast.

The result is a simple web app that extracts color palettes from images and clearly demonstrates the power of modern coding assistants.

What fun projects have you been building with coding agents lately, and how has the experience changed your workflow?