KnowledgeQuizBot

• 

  The main screen

Inspiration

We based our idea in

• Self learning, you can upload your own PDF document and generate training course in the selected language from it and test yourself!
• Exam preparation, you can test yourself before the real experience and notice what topics you need to check again
• Your own style as you can load any PDF you evaluate yourself at any level

What it does

KnowledgeQuizBot generates a training course processes a given PDF document by extracting its text through text recognition mechanisms, it dynamically formulates questions based on the textual content. Users engage in an evaluative interaction by responding to these questions, and the bot employs a feedback mechanism to inform users of the correctness or incorrectness of their answers in relation to the content in the PDF.

How we built it

We used text recognition methods to get the text from the PDF, the library of LangChain to process the text and generate the questions. For the interaction we decided to do a Telegram Bot, on that way you will have this system on the palm of your hand and you can be practicing everywhere!

Challenges we ran into

Learning about APIs and models based on GPT, extract the information from the model and give the desired format was something new for us

Accomplishments that we're proud of

Working in a team where my native language wasn't everyone's first language was both cool and tough at times and have a full understanding of our thoughts using the available tools

What's next for KnowledgeQuizBot

Add statistics allowing you to check your progress and notice where you need to improve, usability creating bullets of information, better feedback system where you can learn what was your mistake

Built With

• python
• telegram