MEDi

MEDi is an AI-driven platform designed to assist visually impaired individuals in efficiently and accurately identifying pharmaceutical products in stores.

To prevent medication misuse, my team and I developed a way to guide pharmaceutical shopping effectively. In February 2025, we launched a platform called MEDi.

MEDi provides step-by-step shopping guide agents by utilizing hand-tracking and object detection rather than broad assistance. For added convenience, we also offer audible recognition features, ensuring users can navigate their shopping experience with ease and confidence.

To optimize efficiency for real-world use, we implemented TensorFlow Lite and MediaPipe, allowing us to perform model quantization and enhance performance on edge devices. The platform is built with Kotlin, ensuring a seamless and responsive experience for potential users.

The MEDi team is constantly striving to improve and develop the platform, setting goals to provide accurate information and a user-friendly UI/UX. We are deeply committed to continually enhancing the environment for people who often remain in the shadows, ensuring they have access to tools that empower their daily lives.

MEDi is currently in its beta testing phase. You can join us by clicking the logo below.

Click our logo!

If you’re interested in our code, check out MEDi’s GitHub repository! 🚀