About me

Intro

I'm a physicist with a strong passion for technology, I'm curious by nature and enjoy learning about any subject. Most of my experience is building products involving AI capabilities, data processing and analysis. I'm also interested in MLOps and cloud practices.
Right now I'm a ML Engineer/Data Scientist for the Products division at NTT Data.

My brief story

BSc in Physics

My BSc final thesis consisted on applying neural networks to find the fundamental state of a quantum many-body system. Basically, I made use of a library named NetKet (https://www.netket.org) to make a simple NN structure, like a RBM or a FFNN, learn the fundamental state of a given quantum many-body system through Monte Carlo sampling data. The plan was to study a couple of different hamiltonians and check how these ML methods can be useful applied to physics, and what we can expect from them.

NTT Data

During my first 2-3 years at NTT Data I built several recommenders, processed lots of documents, profiled lots of users and built lots of microservices. Got to work with ontologies and graph databases, which are quite interesting, wrote tons of ElasticSearch queries, touched some Spacy pipelines and even helped with the custom search engine of the product. Learnt a lot about containerization and DevOps.
On my 3rd-4th year I got involved in a new product idea, which we had to validate, do some PoCs, pivot, and finally build the MVP. There I got lots of experience and more responsabilities on different things as we were a very small team. Got more experienced with DevOps, CI/CD, unsupervised learning, feature stores...
Lately I am also helping in a product that wraps several LLMs and provides a question-answering platform based on corporate documents.

MSc in AI

I studied a very general MSc in artificial intelligence which gave me the opportunity to learn and practice about a wide set of ML fields. Optimization algorithms, supervised and unsupervised learning, deep learning, reinforcement learning, computer vision, NLP...
My final thesis consisted on building an end-to-end upselling and cross-selling system. I got a dataset with information from about 700K customers from an insurance company, and worked on my own to:

• Clean, process and transform the data
• Perform some basic clustering, GIS visualization, time-series analysis...
• Train a cross-selling model
• Train a upselling model
• Deploy and serve both models in the public Cloud

I like

• Photography
• Lifting heavy stuff
• Being healthy
• Freediving
• Diving
• Podcasts
• Nature

Personal projects

I don't have almost any GitHub repositories as most of the code I develop is private, but I will try to build some personal projects for learning purposes. If there's any such thing it should be shown here: