Finding the perfect match with Machine Learning

TestingTime recruits people for usability tests, interviews and other studies for many international clients. Each study has unique requirements and it can be difficult to find people meeting these very specific criteria.
We developed a system that determines suitable candidates out of a pool of over 350’000 people, based on surveys that they have filled out in the past. The system finds matches by analyzing correlations of related questions: For example, it automatically detected connections between income level and preferred car brands. Using such connections the system predicts hobbies, preferences and knowledge. This increases the chances of a successful match to 250%. And by learning continuously from new data it gets smarter day by day.

Key Technologies/Terms

  • .NET/C#
  • ML.NET
  • MongoDB
  • Docker on AWS Elastic Beanstalk

Client

TestingTime recruits test users for usability tests, focus groups, interviews, surveys and more.