“For the penultimate semester of SMC I had always known I wanted to do an internship and gain experience in industry. I was fortunate enough to come across a paper written by Dr Derry Fitzgerald, the CTO of a small company AudioSourceRE and after contacting them, I was lucky enough to be offered an internship. AudioSourceRE is a relatively new company that creates software to separate or demix audio into its constituent components E.g. Split a song into its vocals and accompaniment. It is the culmination of over 18 years of research in audio source separation by Dr Derry Fitzgerald and I feel extremely lucky and privileged to learn under the supervision of someone with such experience in both academia and industry. Even for a company abroad, AAU made everything extremely easy to organise. The process was very easy, and I had no problems relocating to Ireland for the duration of the semester. I really appreciated the freedom given to students by AAU. I work on developing transfer learning models in order to improve the results of the source separation software for specific artists. Every morning, I participate in a 9am round table where there is open discussion of our current project and any organisational tasks. I’m able to see every part of how the day to day running of the company happens and gain invaluable insight into the reality of developing a software product. Two big barriers to entry for machine learning are the large amounts of data and raw computing power required. AudioSourceRE gave me the resources I needed to work on the project whether that was help in finding and developing datasets, AWS GPU instances, or even just help sometimes to understand the state of the art technology I’m attempting to improve upon. Developing the models has been extremely rewarding and I have some very promising results I am hoping to publish in the future. I would never have been able to reach the same quality by myself and I feel like much more than just a student intern here, a real part of the team.”