Praxyk

Praxyk

Machine Learning as a Service

Overview

Motivations :
  • The space of problems that can be solved with machine-learning has drastically increased recently.
  • Large scale data-collection is ubiquitous, even among non-tech businesses.
  • The prevalence of ''big-data'' makes customizable machine-learning tools in high demand.
  • Setting up the infrastructure and configuring it to process one''s data can be costly and time consuming.
  • It often does not make financial sense for smaller businesses to set up their own ML frameworks
(non-new) Idea

Implement the most popular machine-learning tools and offer access to their predictive capabilities as a service through a dedicated API.

  • Abstracts the issue of designing and maintaining a custom ML framework away from the customer.
  • Makes the predictive power of ML more accessible, especially for small businesses.
Services
  • POD - Prediction on Demand A collection of pre-trained ML tools for quick and simple solutions to popular prediction tasks (face recognition, spam detection, speech recognition, etc.). These models are active at all times, ready to respond to task requests.
  • TLP - Templated Learning Platform A managed platform where users can train and configure ML-tool templates. Once configured to their liking, these custom instances can be saved by the user and deployed again at their will.
Project Details

This is a project for CS115 Intro to Software Engineering at UCSC, Fall quarter 2015.

  • John Allard
  • Nicholas Corgan
  • Nikita Sokolnikov
  • Ryan Coley
  • Nick Church
  • Michael Vincent

    This project followed Scrum methodology; relevant Scrum documents can be found in the docs directory.


Updated : 2015-10-01
Created : 2015-09-30