Projects

Resume-Parser employs a Transformer-based Natural Language Processing (NLP) model to streamline the evaluation process, allowing for a quick and insightful glance at resumes. The parser selectively extracts relevant information, eliminating unnecessary content and simplifying the effort required for candidate shortlisting.

This project focuses on sentiment analysis, employing a pre-trained natural language processing (NLP) model to analyze the sentiment. The features of the model include positive/negative analysis, emotion analysis, category of text ( sociology, computer science etc.) and AI generated text classification. The project has a Gradio application which integrates the pre-trained NLP model to process the input data and presents the sentiment analysis results.

Training a transcription model to understand, transcribe and translate hinglish to English, which can have wide applications in note-taking in meetings, classes, etc.

An experiment to develop an EEG-based emotion recognition system using a bandit-based decision-making framework. The aim is to detect emotions like disgust, happiness, frustration, and neutral states from EEG data of participants exposed to various stimuli.

Utilized Reinforcement Learning algorithms to automate stock trading decisions, optimizing portfolio performance.

A project aimed at interpreting American Sign Language to facilitate communication for the differently abled.

We study ML algorithms that can capture a passphrase being typed on laptop based on keypresses heard over a Zoom call. Mentees dived into audio processing and acoustic side-channel attacks and are currently implementing a procedure described in a research paper