After winning the award from the prototype, I went on to develop the first MVP of aMUZE.
I couldn't find a front-end developer, so I had to become proficient with full-stack development. My experience with Swift helped a lot. However, the full-stack tools for web development differ quite a bit.
Before my journey with web development, I must say that I thought that back-end development is supperior to front-end. I was wrong. Front-end development is demanding. Not only because of the plethora of tools that one needs to master, but also because of design decisions that need to be made.
Learning frameworks like React and Express turned out to be a great experience, but with the added complexities and responsibilities of the startup it becaume a big challenge. Except for learning these technologies, I was also doing interviews of potential employees, managing costs, preparing for pitches and meetings with VCs, attending contests, redesigning the pitch deck and leading the back-end development of the project.
It was a lot.
Eventually we got there. aMUZE could generate original lyrics, artwork and music from simple input!
The goal was to create a music-oriented social network where people could not only generate original art and music but also share it with their friends and also import and change music of their friends as well.
I did my best to create an intuitive UI with high quality graphics. I did everything properly. Our virtual machines were protected by a VPN. User files would be stored in our private blob storage. We had a backup strategy in place and monitoring systems too.
For music generation we created our own native solution for aMUZE. The results were not amazing, but they were making musical sense and for simple styles like lullabies and arcade music, were quite good!
For lyrics and artwork we experimented with native solutions as well, but we didn't have much time. Eventually we used of-the-shelf solutions, which with a lot of prompt engineering ended up producing amazing results!
For aMUZE I also implemented a robust authentication system with JWT tokens and a role-based access control. Lastly, I also implemented a notification system with websockets.
Looking back, my approach was wrong. The authentication system especially. I should instead have kept aMUZE open and focused my efforts on the music generation. I was too eager to generate revenue in order to prove to potential investors that aMUZE is a viable business. But my focus should have been on proving that aMUZE is the future and that people will embrace it.
