AI & SMS-based Learning Management System - Introducing EstudIA!

Sebastian Ponce — Thu, 19 Dec 2024 01:38:56 +0000

Access to distance learning should not depend on having access to the latest technologies or a stable internet connection. That's why I created EstudIA: a learning management system (LMS) that works entirely via SMS and MMS. Powered by Cerebras' fast inference system, Twilio, LangChain, and Redis, EstudIA provides an alternative for distance learning in communities where internet access is scarce or non-existent.

How EstudIA Works

At its core, EstudIA transforms SMS-enabled phones into tools for education. Here’s how it functions:

Creating a Class:

Teachers text EstudIA with a subject name and an initial topic to generate study materials and exams using Cerebras-powered AI. A unique class code is created and shared with students.

Joining a Class:

Students enroll by texting the class code. EstudIA then sends them the first reading material via SMS.

Learning and Interaction:

Students receive bite-sized lessons and quizzes. They respond with answers via SMS, and Cerebras provides immediate feedback.

Teachers can dynamically update the class topic or materials, ensuring lessons stay relevant.

Assessment and Feedback:

AI-generated exams test students’ understanding, and grades are sent back via text, making the process seamless and interactive

The Tech Stack

Building EstudIA involved integrating powerful yet accessible tools:

Twilio: Handles SMS/MMS communication, enabling reliable two-way interactions between students and teachers.

Cerebras + LangChain: Generates study materials and exams tailored to each topic, ensuring quality content.

Redis: Manages real-time data, tracking student progress and caching frequently accessed content for fast responses. I also used Redis because I appreciated learning something new, and it was very easy to use, which is always nice.

Thank you Cerebras!

Cerebras has been an incredible enabler for this project. This PoC has been created with their fast inference system as the core for the study material generation. It's been truly awesome and seamless to implement their technologies during EstudIA's initial development stage.

Do it in Keras! - Hacktoberfest

Sebastian Ponce — Sun, 03 Oct 2021 22:09:39 +0000

Modern deep learning architectures and tasks, all implemented in Keras

Have you ever been in a situation where you scroll through forums, reddit, journal posts or even 5 year old github repos trying to find an example for a Deep Learning task that you want to implement using Python and Keras?

Is your code not working because the original code was outdated?

Are you tired of always seeing the same benchmark datasets without the possibility of implementing your own?

Do you like to contribute to the spread of Deep Learning and want to implement your own examples in Keras?

Well, then Do it in Keras is for you!

What is Do it in Keras?

Do it in Keras is a collection of easy to use Jupyter Notebooks hosted in Github which have various implementations of Deep Learning with custom datasets and future exercises for you to learn the most popular Deep Learning library: Keras.

How do I contribute to it?

Do it in Keras is now participating in Hacktoberfest, so you can not only learn a lot more about Keras and Open Source, but you can also win a lot of cool prizes!

If you've never heard about Hacktoberfest, you can check it out here.

To contribute to Do it in Keras, you only need to make a pull request on the official Do it in Keras repository. Anything related to Deep Learning with Keras helps, and you can also upload your own implementation of a Deep Learning task.

DEV Community: Sebastian Ponce