Backpropagation Is Just Dynamic Programming (I Animated It to Prove It)I animated one training step to show backprop's value-reuse — the same trick as dynamic programming. A React + Framer Motion build log.4 min read·algocanvas.hashnode.dev1Jul 1
I Visualized Why AI Models Fail to Train — It's One NumberSame algorithm, same start, one number changed — three outcomes. A visual build-log of gradient descent and the learning rate.4 min read·algocanvas.hashnode.dev1Jun 23
I Built a Neural Network You Can Watch Train — Forward Pass, Loss, and Backprop, AnimatedAn animated neural network that trains in front of you — forward pass, loss, and backprop visualized with React and Framer Motion.4 min read·algocanvas.hashnode.dev1Jun 20