Resume

I'm currently working on efficient machine learning methods for dynamical systems agent-environment interactions. In particular I'm focusing on:
-) Diffusion techniques for object surface recognition in vision-based tactile sensors
-) Computer Vision for real-time action recognition

Ph.D. thesis
Efficient machine learning with resource constraints

Analysis of methodologies to alleviate machine learning computational costs. Focused on: transfer learning approaches, efficient unsupervised methods in plankton monitoring context, and methodologies to compare image datasets as data distributions.

Master thesis
Adapting a recurrent network to tonal analysis of audio files

A study on techniques, methodologies, and models associated with speech recognition to predict sequences of chords played within audio tracks.

Bachelor thesis
Deep Learning Model for graphics accelerators

An overview over convolutional deep learning models. The results were presented as demo during the Nvidia’s GPU Europe technology conference 2016

An internship about deep learning algorithms for CAPTCHA image analysis. The aim was produce recommendations for creating stronger CAPTCHAs and expanded our understanding of applying deep learning to malicious code detection.