In March 2021, we launched a 5-week course that gives an overview of the crowdsourcing approach and has a hands-on practical component. At its core are crowdsourcing projects for ML tasks, and all the materials and practice sessions are based on real industry cases. The participants watch video lectures, take quizzes, and run a full-cycle crowdsourcing project. All the aspects of crowdsourcing are supported with examples from the Toloka team of crowd solution architects, who describe pipelines that are based on real, large-scale industry use cases.

In 2021, Toloka Academy organized a course for CUSO which unites 5 top Swiss universities: University of Freiburg, University of Geneve, UNIL (University of Lausanne), UNINE (University of Neuchatel) and University of Bern. The 3-week course for PhD students focused on the crowdsourcing methodology and techniques that ensure top-quality data. By the end of the course, the participants created projects based on their academic interest.

Many educators run into serious issues with datasets and projects for their courses, which is why we decided to sponsor student projects. Our collaboration with TU Delft started with a hands-on tutorial demonstrating some real use cases. Next, participants of the Crowd Computing course were given an opportunity to collect and label data on Toloka free of charge. As a result, they presented several brilliant group projects.

We provide educators with learning materials that can be easily included in a course or used independently. For example, our materials are used in the Global UX course at Northeastern University. When we share our materials and expertise with faculty members, we also teach them how to apply our methods in the most efficient way.