Skip to main content


Team Name: TOCA

IDEAS Project Pitch: TOCA solves two problems: first, the lack of well-paying jobs in many areas of less developed countries, and second, the need for cheaper labeled data to train supervised machine learning algorithms.

Our team is building a mobile application for budget smartphone users. Through our app, users are able to make money by performing simple tasks that still require the “human touch”. For example, identifying photos of cars from odd angles. TOCA then packages up this data and sells it to our clients, mostly tech companies, who use it as ‘training data’ for machine learning: basically the data that gets the algorithms started. Finally, we pass the money on to our users through small mobile money payments.

The work is similar to Amazon Mechanical Turk, but it does not require a laptop to work - only an Android phone with internet connection. This significantly cuts the cost of labor, one of the biggest bottlenecks in supervised machine learning models.

Sectors: Finance & Entrepreneurship

Project Location: Manila, Philippines