Pranav Barot

Background

About Pranav Barot

I'm a machine learning and robotics engineer with experience in perception, deep learning, digital signal processing, humanoid robotics and controls in research and industrial environments. I am lucky to have the had the opportunity to complete several ML internships as a part of the coop program at University of Waterloo, where I contributed closely to innovative R&D efforts at companies such as Dialpad, Autodesk and Nuralogix. I'm also proud to have completed my Master's at UW, where I worked on developing a full human-robot interaction framework for the REEM-C Humanoid based on acoustical signal processing, computer vision, and control systems based on real-life human behaviour. Some of my past work- I've built computer vision, image and video processing applications (object detection, facial detection, semantic segmentation) for AI based healthcare and robotics systems - I've implemented novel controls systems and sensor fusion for real-time tracking and state estimation on humanoid robots (dynamic modelling based on Vicon motion capture, PID techniques, Kalman and Extended Kalman filters)- I have worked on machine learning for NLP, time-series forecasting, model optimization and feature extraction (text classifiers, ensembled models, Bayesian hyperparameter search, autoencoder based feature extraction and more)- I've performed research in audio DSP and deep learning approaches for real-time systems (DoA estimation, acoustic event detection, VAD, source separation) My research work has been published at conferences, such as ARSO and Humanoids, and journals such as PLOS ONE. Thanks for viewing my profile. Please reach out and connect!