-
-
Our general interface with users
-
client server proess
Inspiration
As engineering students, we receive regular safety training in the labs. We found that current training methods may lack interactivity and engagement. Also, we thought traditional human training was costly. This issue inspired us to generate a way to increase user interaction and reduce costs.
What it does
Our project is a safety training app. The user selects one category, and the app will generate relative questions tailored for engineering professionals to test and give final answers and instructions. It provides them with a comprehensive platform to undergo safety training with no associated costs.
How we built it
We used Gemini as the Al structure to analyze the relevant knowledge in a specific category and implemented the algorithm using Python. The frontend was built by Figma, which gives the final user interface.
Challenges we ran into
One challenge is that when we combined the frontend and the backend, the questions' output was not in JSON format, which Figma cannot read. We solved this issue by reorganizing the .join code in Python. Another challenge is that we did different parts on different computers, which made the testing difficult. The solution is that we used the liked source in VS code and changed the project ID for each person, which ensured we were on the same page.
Accomplishments that we're proud of
We are proud that we successfully utilized the Gemini source provided by Google Cloud. By conducting and modifying our algorithm, we can flexibly insert the Al module into places we want, which greatly helped implement our function goal.
What we learned
Our best gain was learning how to integrate the AI module with real-life applications rather than the single-language modules for chat. We also learned specific techniques, like installing and using libraries provided by Google, and user interface (UI) development.
What's next
Based on our current app, we can improve the feedback messages. This can involve a detailed analysis of the user's response and generating questions based on its weaknesses.
Log in or sign up for Devpost to join the conversation.