I'm Mohammad Abuzar,
Computer Science
University at Buffalo

Being a passionate and focused graduate
student, seeking Research and Development opportunities, I believe that
there is a lot to unearth in the field of Machine Learning coupled with Natural Language Processing which can benefit the society.

About Me

I believe in perfectionism. Apart from being an active member of sports fraternity, and a traveler, I am an optimum result-oriented software development professional bringing experience in software design, development and integration with advanced knowledge of Java and C#, who implements Agile and object-oriented approaches in application development. Here is my Resume


Natural Language Generation for Social Bots
This is an ongoing project, where the idea is to build a social bot which is able to converse with humans coherently. We are trying to leverage the features of Long Short-Term Memory Neural Network, to produce grammatically correct speech output.

Remote Consultant Management
This project was built in Hewlett Packard Enterprise, which aimed to streamline the process of hiring, financing and analyzing technical work force. The team designed an intelligent system to auto assign tasks based on available customer requirements, and availability of an engineer. The system also aimed to proactively predict future failures, of a system in the network, using predictive analysis, thereby increasing the cost effectiveness.

Comparison of search algorithms using inverted indexes
Examines the performance of two inverted indexing similarity search algorithms. The task is to examine and compare two different algorithms: term-at-the-time (TAAT) and document-at-the-time (DAAT). Both algorithms implement a top-k ranked retrieval, i.e. given a query, retrieve the k closest ads sorted by their similarity score.

LeToR using Linear Regression
The goal of this project is to use machine learning to solve a problem that arises in Information Retrieval, one known as the Learning to Rank (LeToR) problem. We formulate this as a problem of linear regression where we map an input vector x to a real-valued scalar target y(x;w). Although the training target values are discrete we use linear regression to obtain real values which is more useful for ranking (avoids collision into only three possible values).

Current Research

Syntax generation using Deep Neural Networks
When two people talk they usually converse not only on the basis of linguistic content uttered but also on the basis of unspoken entailment of the previous sentence. To achieve the goal of such a coherent conversation in a robot we are leveraging the abilities of Recursive Neural Networks. My research mainly focuses on achieving better efficiency while utilizing the intelligence embed in these Neural Networks using Short-Term Memory.

My Research Team
I am conducting research under the guidance of Assistant Prof. John K Pate, along with two team mates.

Assistant Prof. John K Pate
Assistant professor of Linguistics, The State University of New York, University at Buffalo.

Mihir Chauhan
M.S. in Computer Science, University at Buffalo.

Jayashree Chandrasekaran
M.S. in Computer Science, University at Buffalo.


Software Developer - Hewlett Packard Enterprise
As a software developer I worked with cutting edge Microsoft technologies to develop enterprise web applications. I also researched on Automation in Operational Support Systems, where it was required to design a system which robotizes the task of Data Center Heath Checks and Monitoring and provides a detailed analysis and prediction of future downtime.

Freelance Developer
As a freelance developer I developed more than ten prototype robots using various algorithms including PID, SLAM and SWARM. The purpose of prototypes encompassed a wide range of applications including Surveillance Flying Robot, Autonomous wheel chair, Segway and Insect based robots.

Robotics Trainer
I had various opportunities to train professors and their students on building robot locomotion and designing the code to optimize it's movement.


You can reach me via email at abuzarshaikh7 [at] gmail.com