Hi! Any hackers working on ML/computer vision projects?

Hi, I have been a lurker on indie hackers for about a year, and have semi-started my own journey now :)

Are there any indie hackers out there building ML tools or services, who have their own dedicated ML training/inference computers? I'd love to chat with you and learn more about what you are working on.

In my case, I was building a webapp that is related to image/video classification, and I currently run my model inference on my own computer. I have a decent GPU to run training and inference with.

If I want the webapp to return fast results for the user, I thought I would need to get a GPU instance from AWS. I thought to myself, Are there alternatives to this approach, because the pricing with cloud providers for a gpu instance is quite high? Especially for what is to be an MVP that may not actually be profitable? Also, moving the model to a GPU instance would just take more time. The build process for this should be fast, easy, and cheap.

I thought about this 2 years ago and since then I have been developing a service/sdk for building fast computer vision web apps on your own home server. The value is that it allows you to quickly develop an MVP without having to move your model to a GPU instance and paying the price for that.

You can build your MVP on your own home GPU inference backend server that you leave running to serve client requests, the service will generate an API that you can use in your frontend (that can also store user inputted images/videos) and forward user requests (along with the media) to your home server. No need to setup your computer as an actual web server, the service handles all of that for you.

Please let me know if this interests you. I'd love to chat and learn about what you are working on, whether or not it is ML/CV related!

Excited to join the community,