The problem DeployAI.co solves

Heroku for ML/AI developers. Just upload your model if you have any, or simple build it in our platform. And once you create one we will give you APIs to interect with directly, without any worry about scalling.
We have two user flows one for those who are good at coding and know a bit about ML/AI. They may upload their trained model in our platform and we will provide them with our wrapped APIs keeping in mind about the scalibility.
Those who just have the data may select use our GUI tool to upload the data and upload it in a very simple manner i.e. upload all the images of cat together and likewise. And they may use our APIs as well for building any application over the provided APIs such as mobile applications, Raspberry Pi etc.
Our platform enables the researchers and developers to test and scale their ML research work without worrying abouth the things they should not be worried at.
We believe that making a model for an ML project is only 5% of the whole pipeline project. And it is our step to solve that problem by using DevOps and similar technologies combining with the cloud computing capabilities.
[Instruction to run the code in the respective repopsitories]

Challenges we ran into

There is a very less interpolability. Also we had thought that we would complete the first part of our project that is only deploying the existing models and providing them with the APIs in the initial few hours but we ran into several problems due to version in tensorflow.