Implemented a serverless architecture using Cloudflare Workers and the Hono library, reducing infrastructure costs by 50% compared to traditional server setups.
Designed and implemented database schemas using PostgreSQL, applying relational database concepts for efficient data management.
Integrated Arduino IDE for hardware programming to facilitate communication between sensors and the backend system.
Developed a prototype for detecting potholes on roads using an ultrasonic sensor installed on streetlights, providing real-time data on pothole locations to improve road maintenance and safety.
Optimized TensorFlow Lite model integration, achieving 20% faster inference compared to the standard TensorFlow model.
Collaborated with a cross-functional team to ensure seamless integration of the TensorFlow Lite model, contributing to a 98% accuracy rate in pneumonia detection.
Developed and deployed a Flask-based web application integrating a TensorFlow Lite model for real-time image classification.
Contributed to improving diagnostic capabilities for healthcare professionals by providing a reliable and accurate pneumonia detection system.
Focused on creating a minimalist and user-friendly interface to enhance task management efficiency.
Implemented JWT-based authentication to ensure secure access to the application.
Incorporated studies highlighting productivity increases ranging from 10% to 20% through the use of to-do lists, validating the project’s effectiveness in enhancing task management efficiency.