Although I also know my way around frontend development and design as well, I'm far less skilled at it. As such, this website is trying to impress in a different way:
It implements parts of the HTML parser spec to tokenize and highlight its own source code.
You can find the full source code of the site here. It's licensed under the BSD 2-Clause "Simplified" License, so feel free to use the code as a reference to implement your own parser! If you have something interesting to share or just want to tell how the code might have helped you, make sure to send me a message :^)