The search for experts among highly qualified migrations is a crucial issue for developing countries. This article presents a deep reinforcement learning method to address this issue from search engine results on the web. The obtained results can be used by migration sociologists to better understand the diasporas of knowledge, as well as for developing countries to locate their trained experts abroad. Our method is based on queries to search engines to retrieve information about the institution and year of affiliation of each expert. The goal of our work is to define a smart browser capable of assisting this search by generating and observing as few automated queries as possible. This is done by means of a Deep-Q network and various neural network architectures for the Q-value function approximation.