In software engineering, students easily find internships in companies while still studying. To combine their studies and employment, many of them seek to compose their final theses in an industry context, for the benefit of the employer as well as to simplify their context switching between job and studies. This can put the student between a rock and a hard place, as on one hand the employer has certain expectations in terms of working for the company, whereas the supervising professor needs to follow the university guidelines. An additional aspect worth considering is the university as an administrative home for the thesis and owner of the thesis process. In this paper, we study how the different stakeholders – the student, the supervising professor, and the company – should act for the best possible results, so that the company problem gets solved, and the results can be reported in accordance with the best academic practices. The research builds on authors’ collective supervision experience, covering more than 1000 theses (mainly master’s level) and close to a sum of hundred years. The thesis has been mainly supervised in two universities, with the clear majority executed in this setup, but there are also several exceptions where the thesis has been eventually accepted in some other university. The results are expressed in the form of anti-patterns, which consist of a definition of symptoms of a problem, its root causes, and proposals to salvage the situation in a practical fashion.