Maitrey Mehta

Maitrey Mehta

Natural Language Processing and Machine Learning Enthusiast

Contact Me

About Me

I am a graduate student at University of Utah pursuing Masters in Computing (Data Management and Analysis). I graduated with a Bachelor of Technology in Information and Communication Technology from School of Engineering and Applied Science, Ahmedabad University, Ahmedabad. My research interests include Natural Language Processing, Machine Learning, Deep Learning, Text Mining and Information Retrieval and am currently working on Natural Language Generation and its use in recommendation systems.

Soccer is my definition of relaxation. I have had my golden days with chess. I may be a great partner to go to a quiz with.

Courses and Certificates

Major Projects

Natural Language Generation

project name
Research

A research project that deals with generating understandable natural language statements using Deep Learning techniques such as LSTM RNNs, variants of Generative Adversarial Networks etc. As an extension to this project, we try to generate discrete yet correlated data using Autoencoders and Generative Adversarial Networks

View Project
project name

Text Semantic Similarity Open Source

A model to measure the semantic similarity between sentences. Central to this approach is a corpus-based unsupervised learning model based on Second-Order Pointwise Mutual Information. The model relies on commonality of neighbouring trigrams along with TF/IDF factor to build a statistical measure to calculate the semantic similarity score. The syntactic similarity score is calculated using a combined subsequence matching measure.

Find out more

project name

Distributed Secure and Highly Available Cloud Object Storage

A project that aims to provide distributed, decentralized and scalable storage service solution with the help of remote storage nodes provided by connected peers and parallelized worker node provided by the cloud/ master server. The objective of this IaaS project is to enable the user to lease out unused storage and in turn store files in an encrypted and distributed fashion.

Other Projects

Facial Expression Detection in Static Images

An algorithm to detect six distinct expressions using convoluted neural networks. The algorithm is built using a series of convolutions using the feature kernels and maxpooling for image reduction.

Context Switching Simulation Open Source

An Operating Systems Project, implementing a simulation of the context storing, switching and retrieving context as per dispatcher routine as well as testing different scheduling algorithms.

View on GitHub

Voice Controlled Transcriptor

An Embedded Systems Design Project, it is a device that transcribes human speech on paper via a motor pulley system. The speech processing module is developed in Python using Google speech API. The hardware consists of an ATMega 32 microprocessor and a two motor system.

Work Experience

Research & Development Intern - Hirevalley (May 2017 - Present)

The project involves building a job recommender system to a candidate and a candidate recommender system to a recruiter. The platform is a SaaS enabling multi-tenancy. Machine learning algorithms and information retrieval techniques are employed to build the underlying system.

Publications

  • Patel, S., Kakadiya, A., Mehta, M., Derasari, R., Patel, R., and Gandhi, R. (2018). "Correlated data generation using gan and its application for skill recommendation" in 2nd Workshop on Data Science for Human Capital Management collocated with ECML-PKDD