Amelia Villegas Morcillo

Developed generative models for AI-assisted protein design and drug discovery, introducing guided diffusion for antibody sequence–structure co-design and discrete diffusion for all-atom protein sequence generation; secured a ZonMw grant and supervised MSc/PhD students.

Developed and optimized AlphaFold-based models for protein structure prediction, built training and validation pipelines for ManyFold, and integrated protein language model and MSA features to improve model performance.

Developed automatic function prediction techniques for proteins contributing to the CAFA-3 challenge and evaluated hand-crafted versus learned protein embeddings for function prediction.

Developed deep-learning based automatic speech recognition techniques and prepared datasets and scripts for Kaldi to support open-source ASR toolchains.

Grade: cum laude

Activities and societies: Member of SigMAT research group; publications on protein fold recognition, hyperspherical embeddings, and protein language model analyses.

PhD in Computer Science with a thesis on automatic protein fold identification using sequence representations and deep learning techniques; research included protein folding, embeddings, and function prediction.

Activities and societies: Received best academic record award in the Bioengineering specialization.

Master of Telecommunications Engineering with specialization in Bioengineering; thesis on language recognition system optimization combining phonotactic and acoustic information with deep neural networks.

Activities and societies: Received best bachelor thesis award in Signal Theory and Communications area.

Bachelor in Telecommunication Technologies Engineering with a focus on telecommunication systems and signal processing; thesis on protein classification based on physicochemical amino acid indices.