RapID LessonDesigner

Inspiration

Developing Lesson Plans is an essential part of being an Educator. But it is often tedious and time consuming.

As an adult educator who has been using GPT since 2021 in my work, I have experienced firsthand the transformative impact of Generative AI on my work.

After ChatGPT was launched in Nov 2022, I was moved to help other Instructional Designers, Educators, and Trainers to use Generative AI to improve the quality of their learning design.

However, I have realised that many educators still continue to face challenges in crafting effective lesson plans with ChatGPT and other LLMs.

The app I created on Amazon Bedrock is designed to significantly enhance the quality of educators' lesson plans while drastically reducing the time required to create them.

What it does

Instructional Design often synthesizes the art and science of learning into a tangible output, the Lesson Plan.

Instructional Design models, such as Gagne's Nine Events of Instruction, are incredibly helpful in developing impactful learning experiences and effective lesson plans.

However, implementing these instructional design models can be challenging and time consuming. RapID LessonDesigner simplifies the process of incorporating Instructional Design models, such as Gagne's Nine Events of Instruction, into the process of developing a lesson plan.

Users input key parameters and the app generates a comprehensive lesson plan, ensuring alignment and coherence among the various components.

Additionally, the app creates two versions of the lesson plan simultaneously using different language models, providing educators with alternative activities to consider. This feature adds flexibility and variety to the lesson planning process.

Challenges I ran into

Creating and aligning the prompts throughout the app that were required to control the output format, align the learning activities to the topic, lesson objectives, and other Gagne Events took some work.

From a technical perspective, I originally developed and tested the app with Jurassic Ultra as one of the LLMs used. This, however, often resulted in various errors and prevented the app from completing in a reasonable amount of time and often required manual intervention. After some trial and error, Llama-70B was decided on for the second LLM.

Accomplishments that I'm proud of

Creating prompts that developed effective learning activities that are also aligned to the intent of the Instructional Model used was less straightforward than initially anticipated. Working through various iterations and sequences of various prompts was fun and rewarding when things eventually clicked.

What I learned

Apart from any new technical insights and learnings in prompting, one key learning for me was how important it was to have platforms that democratized the app development process and allowed people to develop GenAI apps easily, as opposed to being constrained by the form factor of a chatbot, or the need to have coding skills. The profound impact this could have on society should not be underestimated.

What's next for RapID LessonDesigner

Gagne's Nine Events is just the first ID model to be used in this app. Others such as 5E, Kolb's Experiential Learning Cycle, 4C/ID, and others are planned!

Built With

• amazon-bedrock
• claude
• llama-70b