About Me

Viraj Mavani

I am a software developer and machine learning enthusiast. I am currently at The University of Texas at Dallas working towards my Masters degree in Computer Science graduating in May 2020. I am open to new and exciting career opportunities.

My Career

7-Eleven Inc.

Successfully delivered a trend decomposition system using Facebook Prophet to production with 5.36% mean error. Deployed a React.js web dashboard for viewing the results with backend scripting in Node.js. Skills Utilized - Python, Spark, Facebook Prophet, Scikit-learn, Databricks, React.js, Node.js.

May 2019 - August 2019
Machine Learning Intern

SeRViCE Lab, The University of Texas at Dallas

Designed and developed an image classifier for loop-closure detection for Visual SLAM in an autonomous vehicle setting. Ideated and implemented a new system to make datasets hosted on web servers and a REST API to access them while training thus enabling dataset modifications after release. Skills Utilized - Python, PyTorch, urllib.

Sept. 2018 - May 2019
Research Assistant

Meditab Software Inc.

Designed a computer vision based system for Traffic Surveillance

Mar. 2018 - Jun. 2018
Software Engineering Intern (Artificial Intelligence)

Indian Institute of Technology, Gandhinagar

Conducted research on impacts of cognitive data on facial expression recognition by deep learning classifiers. Resulted in publications at ICCV, CVPR. Skills Utilized - Python, Tensorflow, PyTorch, NVIDIA DIGITS.

May. 2017 - Jun. 2018
Research Intern

My Education

The University of Texas at Dallas

Aug. 2018 - Aug. 2020
M.S. in Computer Science

L. D. College of Engineering

Aug. 2014 - Jun. 2018
B.Engg. in Electronics and Communications Engineering

My Skills

My Projects


A Semi Automatic Image Annotation Tool which helps you in annotating images by suggesting you annotations for 80 object classes using a pre-trained model.

...   ...


A simple RDBMS implemented from scratch with C++17.

...   ...

Intersection Detection from Raster Maps

This project proposes a new method to detect intersections in real time along with training a classifier from Google StreetView Images.

My Publications

SAF-BAGE - Salient Approach for Facial Soft-Biometric Classification - Age, Gender and Facial Expression

Ayesha Gurnani, Kenil Shah, Vandit Gajjar, Viraj Mavani, Yash Khandhediya (Authors Contributed Equally), WACV 2019

ViSHuD - Using Visual Saliency to Improve Human Detection with Convolutional Neural Networks

Vandit Gajjar, Yash Khandhediya, Ayesha Gurnani, Viraj Mavani, Mehul Raval, CVPR 2018

Facial Expression Recognition using Visual Saliency and Deep Learning

Viraj Mavani, Krishna P Miyapuram, Shanmuganathan Raman, ICCV 2017