Full-stack developer · Machine learning enthusiast · Problem solver
About
I'm Dipu — a developer and CS student from Nepal. I build web applications and explore machine learning, with a focus on writing software that is practical, readable, and well-considered.
My recent work spans from Java Swing desktop tools to deep learning pipelines for medical imaging. I'm comfortable across the stack — from database schema to browser-rendered UI.
I learn by building. Every project in this portfolio was started because I wanted to understand something better — and finished because I wanted it to actually work.
Technical Skills
Selected Work
Human Facial Emotion Detection
Real-time emotion classifier using a webcam feed. TensorFlow model detects faces via OpenCV and labels expressions — Happy, Sad, Angry, Surprised — live in the browser.
github.com/Dipu021/Human-Facial-Emotion-DetectionPneumoScan AI — Pneumonia Detection
End-to-end deep learning pipeline that classifies chest X-rays as Normal or Pneumonia. Fine-tuned ResNet-18 on PyTorch, deployed with Streamlit. ~95% accuracy, ~96% recall. Built with 3 collaborators.
github.com/Dipu021/Pneumonia-DetectionUniversity Management System Tool
Java Swing desktop app for university administration — manage students, teachers, leaves, and fee structure with MySQL database connectivity and a clean GUI interface.
github.com/Dipu021/University_Management_System_ToolBilling System
Automates invoice generation, payment tracking, tax computation, and financial reporting. Reduces manual errors and gives businesses a clean interface for customer and billing management.
Online Voting System
Secure electronic voting platform with voter authentication, encrypted data transmission, and real-time vote counting — built to reduce fraud and bring transparency to digital elections.
Contact
I'm open to freelance work, internships, and collaborations. Whether you have a project in mind or just want to say hello — reach out.