expand_less

08

CASE STUDY

Streamlined Matchmaking and Improved Business Outcomes through Machine Learning

CLIENT

Dubai-based technology firm entering
the education sector

The Problem Statement

  • 1

    Entry into a new vertical required a strong start

  • 2

    In the UAE, schools are rated on various parameters in an annual
    government report

  • 3

    Addressing the shortcomings reported is important for schools to build their reputation

  • 4

    However, identifying the right EdTech service partners to fill these gaps was often difficult, as the industry spoke a different vocabulary while promoting their products

  • 5

    Our client was looking to bridge this gap with a comprehensive
    marketplace that would match schools and service providers
    automatically

Project Overview

The Unique
Challenge

  • 1

    Schools in the UAE are made aware of areas for improvement in terms of quality of education, infrastructure, teaching standards and other factors on an annual basis through a governmental report

  • 2

    This annual report is a major consideration for parents and students while making admissions decisions, meaning that schools consciously work to address the highlighted improvement areas to improve their ratings in the annual report

  • 3

    However, with a large number of EdTech service providers in the market, and each offering a wide range of solutions, manually identifying the right solution from the right service provider was a major resource drain for schools

In a nutshell, we needed to set up an EdTech marketplace to facilitate accurate and automated matching of schools’ requirements with service providers’ solutions.

The Approach
  • 1

    Google Cloud’s AutoML was used to build a custom machine learning model atop the client’s proprietary language framework, to facilitate automated, fast and accurate matching between schools and educational service providers

  • 2

    Schools registering on the portal are administered an assessment questionnaire to capture details about their specific requirements

  • 3

    Educational service providers registering on the portal can enter detailed descriptions about the solutions that they offer

  • 4

    A custom ML algorithm was developed which would then perform the matching between requirements and services offered

  • 5

    Training of the ML algorithm was performed based on solution descriptions from the providers and will continue on an ongoing basis

  • 6

    The trained algorithm identifies solutions that map closely to the submitted needs and matches them automatically

  • 7

    The accuracy level 2 months after go-live has already hit 97.5% - over a period of time, the accuracy will approach 100%

  • 8

    The algorithm also understands the solution descriptions submitted, and automatically moderates whether to list them on the portal or not

  • 9

    The portal includes a meeting platform for meetings to be conducted between schools and providers

  • 10

    Providers are also allowed to run campaigns within the platform, targeted to relevant schools via email, without disclosing the specific schools

Architecture

Business
Requirements and Drivers

The architecture chosen for this particular application took careful consideration of the business requirements and constraints detailed by the client.

  • 1

    The use of ML for automated matching offers a level of speed and accuracy that would be impossible with manual processing or even the use of more conventional technologies.

  • 2

    Google Cloud was chosen as the cloud platform for this solution for it’s powerful ML engine combined with optimal costs for this application

  • 3

    High quality code was ensured at every stage of the software development cycle with code quality checks using SonarQube

  • 4

    Application security was a critical consideration with the application being hosted completely on public domain, and monitoring and detection of security vulnerabilities was achieved through penetration testing using Rapid 7

  • 5

    The portal helped the client fill a gap in the market by giving schools a fast and accurate way to find the solutions they need, while also providing service providers with a platform for promotion to a highly relevant target audience

  • 6

    The high level of automation on the platform makes it a highly scalable new line of business for our client, requiring very little manual intervention or staffing

  • 7

    This degree of scalability also makes it possible to expand the portal to address other geographical markets

Simply put, the new EdTech Marketplace built by CloudNow helped the client make a strong start on their core business objective - entering the competitive education sector.

All Case Studies
Solve real-world business challenges and
leverage the power of the cloud

With CloudNow’s Expert Teams in Your Corner.