AI systems, full-stack applications, IoT automation, and research — all built end-to-end with passion.
🤖 AI / ML
2024
An intelligent healthcare assistant providing symptom analysis, disease prediction, and medical guidance for rural communities in Nepal lacking specialist access. Uses NLP models fine-tuned on medical datasets.
📷 Computer Vision
2024
A deep learning computer vision system that detects and classifies facial skin diseases using CNNs. Trained on dermatological image datasets, achieving high accuracy for multiple condition categories.
🤖 IoT / Robotics
2023
A complete smart home system built on ESP32 microcontrollers with real-time device control via Flutter mobile app, voice commands, PIR sensor automation, and Firebase cloud synchronization.
💻 Full Stack
2023
A full-featured student management platform with attendance tracking, grade management, exam scheduling, report generation, and a Flutter mobile app. FastAPI backend with PostgreSQL database.
I'm constantly building new things. Follow my GitHub for the latest projects and experiments.
An intelligent healthcare assistant designed specifically for rural communities in Nepal and similar developing regions lacking specialist medical access. The system provides symptom analysis, preliminary disease prediction, and evidence-based medical guidance via a mobile app that works in low-bandwidth environments.
Millions of people in rural Nepal have no access to doctors or specialists. Basic healthcare decisions are often delayed or wrong due to lack of information. This project aims to bridge that critical gap using AI.
Mobile App UI
Dashboard
Analysis Screen
A deep learning computer vision system that detects and classifies facial skin diseases and conditions. Built with PyTorch and fine-tuned ResNet architectures on dermatological image datasets, the system provides real-time analysis via webcam or image upload.
Input image → preprocessing pipeline → ResNet-50 backbone (fine-tuned) → custom classification head → softmax output with confidence scores. Grad-CAM used for explainability.
Detection UI
Grad-CAM Heatmap
Model Accuracy
A complete smart home automation system built around ESP32 microcontrollers with real-time cloud synchronization using Firebase, a Flutter mobile control app, and multiple sensor integrations for intelligent automation.
A full-featured student management platform for educational institutions. Handles student registration, attendance tracking, grade management, exam scheduling, and generates detailed PDF reports. Includes a cross-platform Flutter mobile app.