Some examples: machine learning scientists and engineers; quants and (quant) developers.

In either case, you're primarily either a developer or a scientist/researcher. Unless you're at an early stage startup or in a principal/managerial role, chances are your time is better spent deeply focusing on one area rather than both. Either putting the system to production or doing research. A lot of the time, these tasks can be carried out in parallel -- while one person is carrying out research, the other can be improving the system.