Projects
Tracking Covid-19 through Social Media using Sentiment Analysis
Worked in team of 7 implementing data pipeline to process unstructured text data from Twitter and Google. Developed data preprocessing and analytics modules to cleanse and gain insight from collected dataset. Developing text classification sentiment analysis deployed on website using Django web framework.
Fraud Detection using Encoder-Decoder Temporal Convolutional Network (ED-TCN)
Implemented deep learning solution to detecting credit card frauds using sequential ED-TCN model. Developed model using Keras achieving a 96% accuracy. Presented project in RowdyHacks Hackathon workshop session with 40 attendees.
Text-based Classification of Questions on Q&A Forums using Bi-LSTM
Developed Bidirectional-LSTM model to detect improper questions using datasets from Q&A forums. Deployed model on website using Flask web framework for users to input their questions for classifications. Developed model using Pytorch achieving a 97% classification accuracy.
Trolls Detector using Bidirectional Encoder Representations from Transformers (BERT)
Implemented deep learning solution to detecting trolls online using sequential BERT model for classification. Built website that uses model to detect trolls on forums websites allowing users to input the posts. Developed model using Pytorch and deployed on website using Flask web framework.
Tutor Tools Software
Built a software application for tutoring services using Java and MySQL. Applied software engineering principles, Agile Method, and software design patter model-view-controller (MVC). Allows tutors and students to manage tutoring sessions and give a visual organization of the collected information