███╗   ██╗ █████╗ ███╗   ██╗██████╗  █████╗ ███╗   ██╗
████╗  ██║██╔══██╗████╗  ██║██╔══██╗██╔══██╗████╗  ██║
██╔██╗ ██║███████║██╔██╗ ██║██║  ██║███████║██╔██╗ ██║
██║╚██╗██║██╔══██║██║╚██╗██║██║  ██║██╔══██║██║╚██╗██║
██║ ╚████║██║  ██║██║ ╚████║██████╔╝██║  ██║██║ ╚████║
╚═╝  ╚═══╝╚═╝  ╚═╝╚═╝  ╚═══╝╚═════╝ ╚═╝  ╚═╝╚═╝  ╚═══╝

Final-year Computer Science & Engineering student at AJIET (CGPA: 8.24) building production-ready systems at the intersection of machine learning, backend engineering, and data science.

I work across the full ML lifecycle — from raw data and model training to inference APIs and deployed web apps. My projects span RAG systems, CNN-based computer vision, ensemble ML models, and real-time Java applications.

Current focus: LLMs · Computer Vision · REST APIs · Data Pipelines

Python · LangChain · FAISS · Google Gemini

RAG-powered document Q&A system with full multimodal support. Converts PDFs and text into vector embeddings via FAISS for semantic retrieval, grounded to source documents.

• Built end-to-end document ingestion, chunking, and embedding pipeline
• Integrated Google Gemini for contextual question answering
• Web-based frontend for document upload and conversational querying

Python · Scikit-learn · XGBoost · LightGBM · CatBoost · FastAPI · Supabase

District-specific crop recommendation engine trained on 24,500+ rows of Karnataka agricultural data.

• Ensemble of Random Forest, XGBoost, LightGBM, CatBoost, ExtraTrees → 89% accuracy
• FastAPI inference API with prediction logging in Supabase (PostgreSQL)
• Designed for real-world agricultural decision support

Python · TensorFlow · CNN · Streamlit

CNN-based image classifier for 38 plant diseases trained on 87,000 labeled leaf images.

• Applied normalization, resizing, and augmentation for robust generalization
• Deployed as a live Streamlit web app for real-time inference from leaf photos

Python · Pandas · Scikit-learn · Power BI

EDA and ML pipeline on 918 clinical records for heart disease risk identification.

• Trained Logistic Regression, Decision Tree, SVM, Naive Bayes, KNN → 86% accuracy
• Built Power BI dashboard visualizing risk across age, gender, and cholesterol levels

Java · Swing · Sockets

Real-time chat application with custom client–server architecture using Java Sockets.

• Built with ServerSocket/Socket APIs and DataInputStream/DataOutputStream
• Dynamic UI with auto-scrolling chat via Swing components and multithreading

Java · Swing

Interactive visualizer demonstrating sorting algorithm behavior in real time.

• Bubble Sort animations with controlled repaint cycles and custom rendering
• Multithreaded comparisons and swaps to illustrate time complexity live