Close

Siddharth Sharma

Graduate Student | Machine Learning | Computer Vision | Web Development

About Me

I'm a Computer Science Graduate student at University at Buffalo. I am passionate about the field of Artificial Intelligence and its application in Computer Vision and Data Science. I'm currently interning at Informatica for the Summer, where I have been working on building front end, backend services as well as Machine learning based log analysis tool. Previously I have worked as a Software Engineer at Sapient Global Markets, India. My primary role there was to design and develop scripts in Java to improve existing functionalities. It helped me learn the technical aspects of software development as well as ethics of corporate work culture.

Experience

Informatica

Software Engineering Intern : Spring Boot, NoSQL, Javascript, sklearn

• Responsible for creating internal dashboard framework to view the performance metrics of the various products.

• Designed the pipeline for gathering data points from regression test cases and build UI for visualization.

• Developed supervised learning based analytics tool for log filtering. The tool achieved 94% accuracy in filtering purposeful logs.

Sapient Corporation

Associate Technology : Java, SQL

• Developed Java based applications for clients following agile methodology. Was involved in gathering requirements, creating technical solutions, performing unit testing and creating documents for deploying to production.

• Developed scripts for importing manual trades into the system. The solution being modular and scalable, brought down the processing time for manual trades by 50%.

• Other responsibilities included working on critical defects, production bug fixing, automating daily activities and improving the efficiency of the system as part of monthly maintenance releases.

Education

University at Buffalo

Aug 2017 - Feb 2019

Master of Science

Vellore Institute of Technology

July 2011 - May 2015

Bachelor of Technology

Projects

Distributed Hash Table

  • A chord based distributed hash table built using android emulators as distributed nodes. All node joins are handled in real time and position in the rings are allocated by avd0.

Weather Web App

  • Build a full stack web app where user can find weather information based on location. Functionalities include storing and indexing weather data, authenticating the user, displaying requested data and filtering it depending on user input.

Digit Classification with Deep Learning

  • Implemented Softmax Regression, Vanilla Neural Network and Convolutional Neural Network models and trained on MNIST dataset.
  • Compared the performance of all models by training and validating them on MNIST dataset. Also proved the ‘No free lunch theorem’ by validated the performance on USPS dataset.

Image Segmentation using Fully Convolutional Networks

  • Trained Fully Convolutional Neural Networks (FCN) for pixel wise predictions using a pre trained VGG net model on PASCAL VOC dataset.
  • Tested the model on PASCAL VOC and MS COCO datasets as well as unlabelled images to assess the performance in real world scenarios.

Human Behaviour Monitoring Using Computer Vision

  • Implemented an algorithm for analysing and predicting human behaviour in videos.
  • Optimized the algorithm to achieve overall efficiency of 80.9% in classifying behaviour and reduced the time complexity of the system.
  • Results were published in International Journal of Engineering and Future Technology. 2016, 9(9), 13-22.

Skills

Achievments

Get in Touch