About
Oracle certified Data Science professional with a strong foundation in Python, TensorFlow, Scikit-learn, and SQL. Skilled in data analysis, model development, and deployment to generate actionable insights for solving real-world problems. Experienced in handling large datasets and building scalable AI solutions.
Experience1
AI/ML Engineer Intern
Jul 2025 — Sep 2025Digital Vint · Internship
Processed a dataset of over 1,000,000 financial transactions to extract and clean relevant features for analysis. Trained and compared multiple machine learning models including Decision Trees and Random Forest to identify the best-performing approach. Developed a Fraud Detection System using Random Forest, engineering 30 key features, and improving model performance to achieve 97.4% accuracy with a 99.36 AUC score. Visualized feature importance and risk indicators, providing actionable insights for fraud prevention and decision-making.
- Processed a dataset of over 1,000,000 financial transactions.
- Achieved 97.4% accuracy with a 99.36 AUC score for the Fraud Detection System.
Projects2
Transformer-Based Social Media Analysis For Proactive Governance · Machine Learning Engineer
Engineered a Multi-Task DistilBERT model for civic issue, sentiment, and urgency classification, achieving 95.30% average accuracy on independently collected real-world Reddit data. Developed a three-stage training pipeline using 19,040 real Reddit posts for domain-adaptive pre-training and 518 civic samples for multi-task fine-tuning, achieving a 0.9286 average Macro F1 score across all three classification tasks. Built an interactive Streamlit governance dashboard with complaint exploration, urgency hotspots, city-level analytics, and LIME-based explainability, achieving 88.33% simultaneous correctness across all three predictions.
- PyTorch
AI-Powered Book Recommendation System
Engineered an intelligent book recommendation system using Google’s Gemini 2.0 Flash Lite model with custom prompt engineering and temperature tuning, achieving structured JSON output with 98%+ parsing accuracy for personalized book suggestions. Developed a robust Python backend with Flask framework, implementing advanced error handling, data validation, and JSON parsing algorithms to process natural language queries and generate contextual book recommendations. Implemented sophisticated data processing pipelines using Python’s JSON library and custom parsing algorithms to extract, validate, and structure AI-generated book recommendations with comprehensive error handling.
- Python
- Flask
- Generative AI
- NLP
Education1
B.E.
2022 — 2026Methodist College Of Engineering & Technology
Artificial Intelligence & Data Science
Certifications3
Oracle - Data Science
2025Oracle
Nvidia - Artificial Intelligence
2025Nvidia
Kaggle - Machine Learning
2024Kaggle