About me

I am a research scientist, specializing in generative machine learning for creative applications.

I bring 5 years of experience in modeling sequential data, including 3 years focused on transformers and nearly a year dedicated to large language models (LLMs). I have trained open source LLMs +1B parameters, primarily focusing on audio-based modeling using RVQ-based neural codecs (e.g., EnCodec). My expertise covers generative modeling with autoregressive and latent-variable models, contrastive and self-supervised learning techniques, optimizing audio and music processing for information retrieval scenarios.

I hold a PhD and a MSc degree in Information and Communication Technologies from Pompeu Fabra University, Barcelona, and two BSc. degrees in computer engineering and industrial engineering from Koc University, Istanbul.