Atlas 🌏🌎🌍

Overview

🚀 Everything you need to build your own AI agent for tutoring! 🚀

🌊 Quick Start Guide

Prerequisites: Python plus Typescript

Check out the package locally

gh repo clone 12Siva/PersonalTutorPOC

Leverage the python virtual environment to bootstrap the dependencies TODO

https://packaging.python.org/en/latest/discussions/wheel-vs-egg/ 

Documentaion 📚

Documentation, tutorials, challenges, and many more resources, visit: docs.atlas.io

🔭 Learning Atlas

📕 Read the docs: https://docs.atlas.org

📚Go through each of the dependencies for this library:

* VirtualEnv: https://docs.python.org/3/library/venv.html
* Django: https://docs.djangoproject.com/en/4.2/
* AWS: https://docs.aws.amazon.com/
* Terraform: https://terraform-docs.io/
* Polychain: https://github.com/polynetwork/docs/blob/master/cosmos/README.md
* ETH wrapper: https://github.com/scaffold-eth/scaffold-eth/blob/matic/README.md
* OpenSea: https://opensea.io/

High Level Overview

System Design

Figure A: In the above simplified system diagram we outline how a user will interact with the LLM model to create a personal tutor agent.

Sequence Diagram

💥 Note: The above diagram is outdated as of 5/9/2023 💥

Objective

Leverage a large language model such as ChatGPT or Google's BARD to act as a tutor.

The tutor agent will be confined to a single subject such as photography. The tutor agent will create a set of topics for the subject.

For example, some topics for the subject of photography will be:

* Composition - creative topic

* Lighting - creative topic

* Lenses - technical topic

The agent will then create reminders to test the knowledge of a previous topic or teach a new topic.

UI/UX Flow

• User inputs what subject they want to learn
• AI will outline what are the main topics of the subject
• User sees main topics of the subject
  • User selects yes/no if the topics are satisfactory

TODO

1. Add illustration of how the memory managment game / Flash Cards will look like

   Step 1: Learn Illustrator - maybe we can leverage Google Spreadsheets for this?

   Step 2: Upload screen mock-ups to github repository

   Step 3: Add them to README.md and to PressRelease.md under the 2nd pivot chapter (crypto-currency section)

2. Add FAQ Section to PressRelease.md to generate the PRFAQ doc

Example Bard conversation given the prompt

Development Resources

Implementation Details

1. Steamship-packages - https://github.com/steamship-packages - Wrapper for the LLM model workflows

   Reference

2. One option for LLM model API is via Google PaLM and MakerSuite Reference

3. We are leveraging the OpenAI LLM Reference

Expected Questions

1. How will students discover this service?

2. Will students return to this service?

   • We plan to launch a companion app / game.
   • Similar to the kids mental math game used with flash cards. We'll reward completion of the game with minted NFTs.

3. What will the minted NFTs look like?

   • Like baseball / Pokemon trading cards

4. How should we charge for this service?

5. What keeps the student engaged with the service?

6. What are the foundational models this is built on top of?

7. How will we mint the NFTs?

Future Work

1. Proof of concept for calling OpenAI API
2. Proof of concept of leveraging GCP's PALM API via Makersuite
3. Proof of concept for minting NFTs on OpenSea
4. Go through to learn corporate finance for Pitchbook: