Posts

I will explain five algorithms for rate limiters with example Python implementations: token bucket, leaky bucket, fixed window counter, sliding window log, and sliding window counter.

I will explain how to understand and implement backpropagation, the mechanism used to train AI models, with computational graphs and intuitions.

Some people still dismiss AI as hype that can only handle simple tasks. I disagree. Here’s how AI has fundamentally changed the way I write, learn, and build.

I built a portfolio analysis tool entirely through vibe coding — 100% of the code was written by AI. Here’s what I made and what I learned.

This post shares my perspectives on why every software engineer should use AI: the productivity gains it enables and the shift it is driving across the industry.

While many investors understand the power of compound returns, they may not fully grasp that compounding can also result in a significant downside. This post explores this double-edged sword.

When nvidia-smi shows free GPUs but PyTorch throws OOM errors, here’s how to find and kill orphaned processes still holding the GPUs.

I will start from the problem of semantic segmentation, introduce how to use CNNs to solve it, and talk about fully convolutional networks, a widely used framework for semantic segmentation, in great detail. Moreover, I will analyze the MXNet implementation of FCNs.