code · create · reflect · repeat
3rd year B.Tech AI & ML student at IGDTUW · CGPA 9.64
I am passionate about solving real word problems with technology and apply my skills by creating projects.
Grade 8 pianist (Trinity College London, Distinction)
Song Popularity Prediction · Code · Live
End-to-end ML pipeline for predicting song popularity. Improved from a 57.7% baseline (Logistic Regression) to 77.5% accuracy with XGBoost, achieving 0.859 ROC-AUC. Feature engineering, model comparison, full evaluation.
Python Pandas XGBoost scikit-learn Matplotlib
Aasha - Conversational Mental Health Chatbot · Code
Custom emotion detection model using keyword analysis and HuggingFace transformers. Prompt-engineered for context-aware, grounded responses.
Python HuggingFace Flask Prompt Engineering
ReadShelf — Mindful Reading Web App · Code
Flask-based full-stack app with personalised reading analytics, a Word Vault using the Dictionary API, and a Pomodoro-style reading timer.
Python Flask JavaScript HTML/CSS SQLite
If you're building something thoughtful or just learning together, let’s talk 🤍
- 💬 GitHub messages always open
Building thoughtful tech, one project at a time ✨