Used state-of-the-art transformer models fine-tuned on legal text data and focused on key tasks such as text classification, text generation, and text summarization using BERT, GPT-2, and PEGASUS. Designed streamlit app where user can upload legal documents and summarize text. Fine-tuned PEGASUS model on indian legal documents for text summarization task. Published the model on Hugging Face. Link to GitHub repository

Designed a Streamlit app to predict the price of used cars based on inputs provided by users like manufacturer name, engine type, and production year. Employed AWS services like AWS Sagemaker for training XGBoost regression model and AWS EC2 instance for streamlit app. Obtained R2 score of 90% for the XGBoost regression model. Link to GitHub repository

Developed 2 Tableau dashboards to highlight the severity of crime and mortality trends in the US that provided an opportunity to inform evidence-based policymaking. Designed stacked bar charts showing victims affected under different age groups. The dataset was collected by the Federal Bureau of Investigation’s Uniform Crime Reporting (UCR) Program through the National Incident-Based Reporting System (NIBRS).