ETH Graduate Student

Yezers is a startup which wants to tackle the most urgent issues in Italian public policies(e.g, education, plastic reduction). In order to do it, Yezers wants to use highly selected young people coming from different backgrounds and using their expertise to make these proposals real.

The scope of this internship was to research about a new solution for the detection of arc fault signal in residential network.- Introduced a suitable Validation Framework - I tried to extrac features from different domains - Implemented an Encoder-Decoder LSTM architecture for anomaly detection.- Finally, I decided to use a novel feature importance framework

The scope of the project at IBT solutions was the development of a software capable of detecting the number of drivers of a car using unsupervised data gathered the black box of the same car. The project was mainly divided into two parts. First of all, from the gathered data I had to obtain useful features which could have helped in detecting different drivers behaviour. Secondly, I developed a model using algorithms for unsupervised learning in order to cluster our data and identify different driving patterns which I assumed to be the number of actual drivers. My supervisor during the project was the professor Carlo Brandolese