Name: Alberto Becerra Tomé
Type: User
Company: Universidad Politécnica de Cataluña
Bio: Data Scientist | Physics and Mathematics Graduate | MSc Artificial Intelligence @ UPC | MSc Big Data and Business Intelligence @ EOI
Location: Spain
Blog: https://www.linkedin.com/in/alberto-becerra-data-scientist-ml-engineer/
Alberto Becerra Tomé's Projects
Automate awarding Open Badges to recipients without the overhead of a server
Exercises for learning the bash command line.
:zap: Dynamically generated stats for your github readmes
Project developed and ran in cluster from Barcelona Supercomputing Center (BSC)
En este proyecto desarrollaremos un Hundir La Flota todos juntos.
Notes and examples from the course Introduction to Quantum Computing (https://qiskit.org/learn/course/introduction-course/)
Train DL Models from Kaggle in CPU (local) VS GPU (AWS EC2)
Project in Spanish [future releases may be in English]. This is a personal NLP project for self-development and amusement inspired in the popular Disney serie "Star Wars: The Mandalorian". A transmission classification subsystem is built as part of the Communications Interception System.
SemEval 12's Semantic Textual Similarity task, specifically in paraphrase detection
Workshop SQL impartido para los alumnos de los Bootcamps de Data Science y Full Stack de The Bridge
Churn is a one of the biggest problem in the telecom industry. Research has shown that the average monthly churn rate among the top 4 wireless carriers in the US is 1.9% - 2%.
This project aims to use the knowledge obtained from Udacity Machine Learning Engineer wit Azure Nanodegree to solve an interesting problem. In this project, two ML models are created: one using AutoML and one customized model whose hyperparameters are tuned using HyperDrive. Performance of both the models are compared and the best performing model is going to be deployed. This project will demonstrate author's ability to use an external dataset in workspace, train a model using the different tools available in the AzureML framework as well as the ability to deploy the model as a web service.
Udacity Machine Learning Engineer Nanodegree Project. Both the Azure ML Studio and the Python SDK will be used in this project. First we start with authentication and then run an Automated ML experiment to deploy the best model. Next, we will enable Application Insight to review important information about the service when consuming the model. And finally, we will create, publish, and interact with a pipeline. Once all of that is complete, we will create a short screencast and a README to demonstrate all your hard work.
In this project, the goal is to create and optimize an ML pipeline. For that aim, it is used a custom-coded model—a standard Scikit-learn Logistic Regression—whose hyperparameters are optimized using HyperDrive. In addition to this, AutoML is used to build and optimize a model on the same dataset, so that it is possible to compare the results of the two methods.
Wordle es un juego de palabras basado en la web, creado y desarrollado por el ingeniero de software galés Josh Wardle, y publicado por The New York Times Company desde 2022. Los jugadores tienen seis intentos de adivinar una palabra de cinco letras, y cada vez que adivinan una palabra reciben una respuesta en forma de fichas de colores que indican si las letras coinciden u ocupan la posición correcta. La mecánica es casi idéntica a la del juego de papel y lápiz Jotto, de 1955, y a la de la franquicia de juegos televisivos Lingo. Wordle tiene una única solución diaria, en la que todos los jugadores intentan adivinar la misma palabra.