Having been passionate all my life about music, I plan on trying to combine my electrical engineering and computer science background with my passion for music. My master's thesis in the field of Computational Creativity consisted in a program that generated musical artifacts inspired by images. I also love teaching and helping others, having tutored other students at a tutoring center in different subjects (Artificial Intelligence, Machine Learning, Signals and Systems, etc.) as well as having been a Graduate Teaching Assistant teaching Digital Systems at Instituto Superior Técnico (IST) for which I received an Excellent Teacher award. I have also worked as a Software Engineer at DocDigitizer. Currently I am a PhD Candidate in Informatics Engineering - more specifically in the fields of Music Information Retrieval and Computational Creativity - at CISUC from the University of Coimbra where I'm studying tapped rhythm interactions and developing a co-creative tool that will help users leverage these interactions. I also have been awarded a Fulbright Scholarship for the school-year of 2023/24 where I did a Research Visit at Carnegie Mellon University (CMU) under the supervision of Prof. Roger B. Dannenberg. In the first stage of my PhD, I experimented with SOTA VAE-based timbre transfer systems, such as RAVE, and explored whether and why they could achieve percussive timbre transfer, including using latent-space visualisation techniques such as t-SNE. Having found out that these systems could not perform zero-shot percussive timbre transfer, I reckoned a more grounded approach was needed, namely by inferring more information from percussive onsets to condition the timbre transfer models in the intended direction. This led me to define a novel classification task, based on Automatic Drum Transcription, called Automatic Drum TapScription (ADTS). To achieve ADTS, given the lack of a paired dataset, I developed a synthetic strategy to build a synthetic dataset, which I used to train several classification models based on both TCN and CNN architectures. Right now, I’m exploring hyper-parameter fine-tuning and transfer learning strategies using Weights and Biases sweeping tools, feature engineering techniques — which have been shown to improve correspondence across different percussive domains — to study their impact on classification models, and data augmentation techniques to improve the models' generalisation to real-world performance.