When the sample of active LinkedIn users ran out, and Habr, GitHub and HeadHunter stopped giving results, we realized that my ideal candidate wasn't actively looking for a job and has been absent from social networks for a long time. It was time to start searching ... research! We literally began to look for studies that were in a similar field to this vacancy. Some of the authors turned out to be professors and other employees of the education system, who weren't our target audience. We imagined our ideal candidate to be ambitious scientist with a background in commercial research. And among the authors of scientific publications there was a considerable percentage of such specialists.