I'm a principal research engineer with Ph.D. degree in Electrical and Computer Engineering. My interests include Machine Learning (ML) applied to Natural Language Processing, Automatic Speech Recognition and Computer Vision.

In addition to some theoretical analysis on ML models, I invented ML algorithms, of which I highlight maximum margin training methods for neural networks, such as maximum-margin GDX (MMGDX) and Support Vector Neural Networks (SVNN), which is based on the regularization method Eigenvalue Decay. I also invented the feature selector based on genetic algorithms and information theory, and the parallel seq2seq model, wich was applied to generative chatbots. You can find most of my publications on Google Scholar, and codes available on GitHub and Mathworks.

I’m currently working on long-term research on Machine Learning and its applications in automatic speech recognition (ASR), including audiovisual ASR (voice + lip reading) and low footprint neural language models at Cerence Inc.