Reinforcement learning, large language models, program synthesis.

Working on a project for the application of model-based approaches to robotics, with approaches inspired by active (Bayesian) inference and stochastic optimal control and with the goal of exploiting uncertainty to deal with more realistic problems.

Teaching got the modules on artificial neural networks and fundamentals of machine learning.

Teaching seminars for modules including Mathematical Concepts, Further Programming, Databases, Computer Vision, Data Science Research Methods.

Research on Approximate Bayesian Inference (variational free energy methods) with applications to control theory/robotics/neuroscience. Working also on the links and extensions of deep learning techniques, especially Restricted Boltzmann Machines/ Helmholtz machines.

For two months part of the EON project. EON (ELSI Origins of life Network) was a project providing space for researchers with interest for theories on the origins of life. In Tokyo, Manuel worked on computational models of pre-biotic chemistries studying properties necessary for the formation of life as we know it and life as it could be. In this project he used a mix of control and information theories, trying to identify measures describing the information flows and storage of living systems once basic homeostatic processes begin to take place. A brief summary of my work can be found at the link provided below.

Portfolio optimization, Reactive search optimization

Management of a computer lab extra activities in a middle school, teaching the basics of computer programming using LOGO language in a fully open-source environment based on Ubuntu (Edubuntu).

Proposing and promoting new activities in the district focusing on youth people.

Clothes and relative accessories assembly.