IEEE RCSM 2025 paper — Evaluating the portability of BERT-based misinformation detection from Twitter to Bluesky. Achieved F1-score of 0.73 in zero-shot transfer, improving to 0.997 after platform-specific fine-tuning.

Hybrid Web + Desktop application for chemical equipment parameter visualisation. Built with Django REST, React frontend, and PyQt5 desktop client with pandas-driven data pipelines.

Novel entropy-based password generation leveraging real-time biological motion patterns of goldfish captured via computer vision. Uses SHA-256 hashing with statistical randomness validation.

Machine learning models to predict leptospirosis risk using climate and epidemiological datasets across multiple European regions. Features spatial risk pattern visualization.

AI Resume Tailor Suite — leverages NLP and AI algorithms for personalized resume optimization with ATS optimization techniques. Led a 75+ member team as Project Task Lead at VITB AI Innovators Hub.

Blood Cell Detection system using YOLOv8 object detection. Trained on annotated microscopy images for real-time identification and classification of blood cells.

Self-hosted text censoring API with dictionary + ML fallback. Built with FastAPI achieving sub-millisecond latency for real-time text moderation.

Data analysis and visualization of Indian Railways ticket booking dataset to identify interesting trends, patterns, and insights using pandas and plotly.

Fact verification tool built with TypeScript for checking the veracity of claims and statements using NLP techniques.

CPU-based Speech to Text Model using Vosk for offline speech recognition. Designed for low-resource environments without GPU dependency.