- 🎓 MSc in CSE @ United International University.
- 🌐 Building full-stack apps with the MERN stack, Next.js, and Django.
- 🧠 Researched on Video Moment Retrieval and multi-modal deep learning.
- 📄 Planning to publish in ML/DL — watch this space.
- 🎮 Game development is on the horizon, one day:D
Programming Languages
Frontend Development
Backend Development
Databases
Auth & Real-time
AI & ML
Tools & Platforms
A full-stack platform where developers can discover projects to join, post ideas seeking teammates, and collaborate through a shared workspace — Kanban board, real-time team chat, notifications, and activity tracking. Think GitHub meets Trello, built for developers who want to find a team and ship together.
📁 Repo (coming soon)
Full-stack restaurant platform with customer ordering and admin analytics. Features menu browsing with search/filter/sort, real-time order tracking, table reservation, and an admin dashboard with revenue analytics, inventory alerts, and role-based access control.
🌐 Live Site · 📁 Client Repo · 📁 Server Repo
Full-stack job board with seeker, employer, and admin roles. Includes a resume builder with live preview + PDF export, application status pipeline, keyword-based job alerts via email, and employer analytics dashboards. Built on Next.js 14.
Connects patients with certified therapists for appointment booking, expert-authored blogs, course enrollment with progress tracking, and secure session-based auth with XSS prevention.
Desktop app with three-role system (customer, mechanic, admin), real-time in-app messaging, tiered car wash booking, parts inventory tracking, and professional PDF invoice generation.
Undergraduate thesis. Built a custom surveillance dataset (200 videos, 622 action classes) to address domain gaps in benchmarks like Charades-STA. Evaluated 6 transformer-based VMR models and designed a multi-modal encoder-decoder integrating video, audio, and text. Deployed on Raspberry Pi 4B with InsightFace + YOLOv8, with a React Native app for real-time moment retrieval.
ML pipeline forecasting SaaS revenue across 2,500+ records (2020–2024). Compared Linear Regression, Random Forest, XGBoost, and SVR with Optuna hyperparameter tuning (30+ trials). Applied SHAP for feature interpretability, evaluated with RMSE, MAE, and MAPE.