recboard - Recommendation models as a service

Recboard aims to provide recommedation models as a service. We use openrec (Check it out, its cool!) for modular recommendation systems.

Objectives:

Recboard plans to satisfy the following objectives:

• Create recommendation systems on the fly via a UI or a command line tool.
• Visualize model training (running locally or in a remote machine) and results.
• Serve the trained model as a REST API.

Contrbutions

This project is not open for contributions, as of now. This is a semester project, for now. Once we have a PoC that resonates well with users and matches expected performance, then we will open this repo for contributions. :)

Wiki

All fun stuff here

Fun questions

Why are we not using golang?

We know that golang would have been a better choice considering better parallelism via goroutines but we are constrained due to human factors (experience, community here, libraries we are using and most importantly time [This is supposed to a PoC, for now.].).

If you are using python, Why no react?

We know react has faster rendering and other benefits. But it has a steep learning curve and we want to complete a user-testable PoC ASAP.

Okay, but atleast you could have used gRPC? You might want to read this.

Other Feel free to roast link or link

Stack we are using

-Backend

• Django (python)
• gRPC

-Frontend