Research on Graph Neural Networks, Generative Deep Learning, and structural optimization, and their application in materials discovery. Use of Docker, Airflow, Kubernetes, Kubeflow...

I learned about:- Machine Learning (the fundamentals and the usage of Python's scikit-learn library).- Deep Learning (fundamentals and hands-on experience with TensorFlow 2 and Keras).- Generative Deep Learning (fundamentals, state-of-the-art methods for Materials Science and Computer Vision).- Machine Learning with Graphs (fundamentals, state-of-the-art methods for Materials Science).- Developer tools (Docker, JupyterHub, git).- Working in a team, following Agile methodologies.

Grade: 9.3/10Title: "Exploration and optimization of the latent space of a generative deep learning model for the inverse design of lithium-ion battery cathodes".Supervised by Roberto Gómez-Espinosa Martín.In this work, I researched, analysed and implemented several disentanglement metrics for representation learning and generative models. Then, I improved a generative deep learning model (variational autoencoder) for the inverse design of inorganic materials, by modifying the model's architecture and performing hyperparameter tuning. Finally, I identified the strengths and weaknesses of the model, and proposed solutions and alternatives.Written and defended in English.

Grade: 9.8/10Title: "Production of the HIV entry complex proteins for functional and structural studies".Supervised by Dr. César Augusto Santiago Hernández and Dr. Mario Mellado García.During this project, I learned and applied molecular biology techniques related to protein expression and purification, cell cultures, and the design of genetic constructs.