Overview

This Site was built with the Astro dev framework

A static site rendering language that utilizes TypeScript and pre processing to deliver ultra fast static content to the client.

Basically, this means you can still do the all of the cool functionality stuff that JavaScript and TypeScript allow you to do, without having to wait for the load time at all! This means near instant content for the user and zero wait time (in contrast to slow wordpress sites)

Deployment

This site is deployed to Github Pages. Why? Simple really. It is a FREE hosting service. So you can utilize unlimited static site cloud hosting as long as you have your code repository on github. You can check out the repo using the link at the end of the page!

Features

Outside of the cool Astro stuff that we have already discussed, there are a few more things that I want to highlight about this site that I find to be cool.
The site uses custom Tailwind CSS, which allows for extensible styling site-wide, and control over every single hex code used.
There is also a custom favicon (the little image in the left corner at the top of your browser tab). Not massively complicated to make, but I thought it was neat learning how to customize it through the HTML head.

I got to use Astro’s new Content Collections feature to programmatically make project pages, without having to go and manually create each new page every time I want to post a project. This is massively useful and I highly recommend you go check out Astro’s stuff if you ever need to code a website!

Learn more about Astro →

I also got to play around with some of the new AI coding tools to help build this project. My university offers GitHub Copilot for free for students. I have been enjoying using agent mode in VS Code Insiders build edition. The models I primarily used to help me out here were the new (as of the time of writing this) Gemini 2.5 Pro 05-06 Preview, and GPT-4.1 from OpenAI.

It really is pretty incredible how much these things can expedite a coding workflow. Using various MCP servers, I can pull in real-time documentation for the tech stack that I am using and do more high-level thinking rather than rote syntax memorization.

Gemini 2.5 Pro is a large reasoning model, so the output is slow, but it can handle a lot of context and do “harder” tasks. I was actually very impressed with GPT-4.1. It’s a tiny fine-tuned model built for coding, and tiny models mean very fast inference speeds, so you get the response back instantly.
It does struggle a bit with cross-file edits (as you would expect for a small parameter model), but it really does exceptionally well with 1-3 file operations (if you are using a programming language that it has been trained on).

Takeaways

Overall this was a very enjoyable project. I look forward to maintaining the site and learning more about web dev.