The problem Build.AI solves

Python is popular amongst data scientists and researchers due to its ease of use. Build.AI is a GUI tool that makes comparing various machine learning models easier. Build.AI helps researchers and novice AI users get the best model for their dataset. It provides a seamless three-click experience or heavy customisation through the user-friendly dashboard. Build.AI can run 20+ models in the backend and display comparative visualisations.

Challenges we ran into

Since it is a utility tool aimed at making comparisons of various models easier, creating the perfect UI for a pleasant user experience was a challenge for us. We wanted to make sure that our tool saves time as well as is easier to use. Therefore, we did a lot of research on friendly UI elements and adopted some of the best practices.

On the backend side of things, one of the major challenges is identifying features and labels in unknown datasets. Various checks have been placed to ensure accurate predictions and evaluation of models.

Another challenge we are facing is the amount of computation power it takes to run all the models. For that, we are aiming to provide users with the option to run the models on their local devices or pay a nominal fee to use our hosted VMs. Thus, researchers with their own GPUs and supercomputers can get quick results by using their own devices.