Machine Learning Marketplace, for Everyone


We want to narrow the gap between the industry and the ML expertise and give talented individuals chances to commercialise their work with simple clicks and earn revenue for their own work. Since the SaaS concept becomes more and more popular, the marketplace of ML engines grows rapidly. However, configuring such platform still needs certain level of knowledge of deploying things to the server. We, as the ML Marketplace, attempts to simplify these steps and integrate these platform together to form a comprehensive utility for Machine Learning.

What it does

  • Upload models
  • Decides prices for the API call
  • Select different models with an inspection of the rating (currently only depends on accuracy) and price

How we built it

  • Backend: RESTful api using Spring + Kotlin
  • Frontend: Material UI + React
  • Database: Firebase - Cloud Firestore
  • Hosted On: Google Cloud Platform, App Engine

Challenges we ran into

  • Could not design a good and modern UI
  • App Engine has size receiving limit
  • Google Cloud Plaform SDK's inconsistent python version -> we have to switch to Kotlin as our backend server

Accomplishments that we're proud of

  • Bring the idea to proof in 24 hours
  • Learn how to prettify front end during the hack
  • having a good idea for hackathon

What we learned

  • Proper way of using GCP
  • How to prettify front end
  • Time management is importants

Different usage of Google Cloud's SDK

What's next for MLM

  1. Better pricing method
  2. Containerize the model
  3. Able to host on client side
  4. Keep different version
  5. Community develop for community
  6. API authentication
  7. Focusing on small business
  8. Review and rating on models
  9. Upvote problems to get developer attention
  10. API management dashboard

Built With

Share this project: