Projects

Developed a Python-based virtual assistant featuring secure facial recognition login using OpenCV, LBPH face recognizer, and Haar cascades, supporting real-time detection and authentication.

Integrated Google Gemini AI for conversational capabilities and pyttsx3 for text-to-speech, enabling seamless user interaction through voice commands and AI-driven responses.

Packaged the entire system as a standalone .exe file on a portable pen drive, allowing users to securely access their personal assistant and chats from any computer without logging into external accounts.

Build a predictive ML system that forecasts employee attrition using historical HR data. Achieved 92% accuracy with XGBoost, identifying critical factors such as job satisfaction, work-life balance, and income level impacting employee turnover. Visualized insights with Seaborn & Matplotlib.

Designed a user-interactive web application using Gradio, enabling HR teams to proactively retain talent, reducing reliance on reactive measures like exit interviews. Deployed on Hugging Face for real-time access.

Engineered a deep learning-based water potability prediction system using ResNet model, achieving 71.8% accuracy and 0.79 AUC. Incorporated SMOTE for class balance and deployed a real-time Gradio web interface.

Incorporated Google Gemini AI for intelligent model insights, enabling natural language feedback and performance diagnostics; contributed to SDG 6 by promoting scalable, low-cost water quality assessment solutions.

Built Salesforce CRM modules (Leads, Campaigns, Cases) and Tableau dashboards simulating Deliveroo’s market re-entry in India, generating insights on customer retention, revenue growth, and competitive positioning.

Performed sentiment analysis and user segmentation on Zomato and Swiggy datasets using TextBlob and Tableau, driving targeted loyalty campaigns and evaluating CRM impact on customer satisfaction.