In 2019, I wanted to make a difference so I embarked on a large passion project to create a service for managing and rewarding extracurricular volunteer work at Innopolis University. I wasn't in it alone, working with two of my mates from university, but it is only when I look back at it that I realize how much was there to be desired in terms of project management. I designed the whole thing from scratch and played a major role in both the frontend and the backend. This was my first experience with a proper JS framework, and it all just clicked.

Also I may have been in love while doing this project, which really helped with the motivation :)

In order to design the Innopoints, I needed a solid UI kit. So I designed my own. Somewhat Material-ish, but at the same time, not at all. From the beginning it was made to be accessible by default, though it still has a way to go until it will get the title of “fully accessible”. My teammate from Innopoints joined me on evolving this UI kit into a separate project. We called it Attractions. Then came the trouble of creating an npm package and all the learning that came with it. The Svelte ecosystem wasn't as rich as it is now, so we wanted to give developers the benefits of strong typing without the burden of committing to TypeScript. I'd say it went well.

A curious front-end developer taps into the unknown — practical machine learning! It's amazing how little effort is required nowadays for a very solid computer vision system, thanks to the state-of-the-art YOLOv4/YOLOv5 model. This project is not about that, however, and what it does surely doesn't require computer vision, but take my word for it, YOLOv4 is insanely powerful.

So what's this project about, then? Well, imagine if a powerful machine learning model of the YOLO variety ran in your browser. No need to wake up, it's not a dream. This project is a proof-of-concept that you can embed a computer vision model right into your SvelteKit application and export it into a bunch of static files to serve on GitHub Pages! This project was a solo effort, and that includes an all-nighter to get the project the best it can be before it has to be presented next morning. Yeah, not particularly proud of the sleep deprivation 💤.

Remember the Innopoints team? Reunited with two other strong developers into what is basically an unstoppable team, we started working on... a podcast player? Not just any podcast player, though, as the mission was to create the most robust podcast player out there.

Tools of the trade? Unit tests, integration tests, end-to-end tests, security tests, accessibility tests, visual tests- *deep breath in* CI/CD pipelines, type checking, uptime monitoring, beautiful architecture, GitHub authentication, unmanaged backend with Supabase *deep breath out*. Impressed yet? Then go check out the README in the repository, it speaks quite well for itself.

This project was a ton of fun, and a way for us to discover just how great it is to build a web app these days. I managed the team and was the lead developer, working on the project from all aspects. Also I created the design, sweet and simple. And a README (my best one yet!)