aka. voidkandy

What began as a love for guitar and music composition became a passion for production and sound design. Ableton was an excellent gateway drug into software. Looking back, it's hard to tell if I was more interested in the music or just pushing the DAW as far as I could. After completing a multimedia music project and spending months grinding self-promotion on social media, I was dissatisfied. Years of releases, combined with an admittedly half-baked marketing strategy, had not resulted in the music career I naively thought it would, something had to change.

I had always been interested in programming, but I believed it was 'too complicated' for the likes of a grug brain like myself. The accessibility that ChatGPT brought to programming was the push I needed to really dive into it. I began writing stupid little Python scripts with the help of GPT-3.5 and eventually found myself fascinated with machine learning models, which led me to moving to a part-time position at my day job to study machine learning on my own.

A few months later, I was grinding through the Fast.ai program led by Jeremy Howard. I had a few classification and fine-tuned models under my belt, and I wanted to go for something a little more ambitious. So, I became an unofficial member of my friend's senior capstone project group. We were to build a speech separation model using a RAG architecture. We failed, but learned a lot, and I had to admit to myself that I had a significant gap in my knowledge of computer science that was holding me back.

Since then, I've taken multiple online computer science courses and decided to learn Rust in the hopes that its complexity would make other languages relatively easier to learn, an assumption which has paid off for the most part. Since then, I've worked on multiple projects and gained experience with multiple languages and technologies.

My introduction to webdev came when I began as an intern for a small startup. I had very little supervision and was thrown into a massive MERN stack codebase. I had never written any Typescript before and was brand new to document-based databases like MongoDB. I started slow but eventually became one of the most productive interns on the team, making and implementing design decisions on the frontend as well as contributing to the payment system on the backend. I'm now comfortable developing with JavaScript/TypeScript, but I'm very partial to using HTMX on the frontend with either Go or Rust on the backend.

I have built my own Rust crate for creating AI Agents and have utilized it on multiple projects. Building, updating, and maintaining the crate has been a great exercise in API and architecture design. I've kept a blog of the process which has been great for documentation and in deepening my understanding of Rust. Because of this crate, Rust has become my language of choice.

My favorite personal project by far has been my AI LSP. Building this has deepened my understanding of my editor (NVim btw), LSPs and even client/server relationships in general. The LSP also uses my AI Agent Rust crate under the hood and has informed a lot of API and quality of life changes to the library. I'm still actively developing the LSP, there is something so satisfying about building a tool that is so broadly applicable.