Nowadays, we often see job ads like “Machine Learning Engineer”, “Computer Vision Engineer”, “Applied Scientist”, etc. There is a high demand for such profiles, not only from the big guns (FAANG), but from medium-small companies as well. Basically, they are looking for somebody able to implement cutting-edge features into new applications, or even develop innovative characteristics from scratch. Needless to say, this job requires a set of skills which are not common for Software Engineers. This presentation analyzes the differences between usual approaches to software development and those required for applied research. We will show some use cases regarding typical scenarios in applied research. Furthermore, we will discuss how good practices for generic software development can be (and should be) successfully implemented for applied research.