Atlas AI

• 

  folder directory

• 

  Task Details Page

• 

  Dashboard/Homescreen

• 

  Database schema

Inspiration

I was mainly inspired by personal necessity. I as a student, work on multiple things a day, and the limited time makes it hard to keep track. And especially learning something new in this age of information is a daunting task. I wanted to simplify it. I wanted an AI-based dynamic to-do list, that would maintain memory of my tasks

What it does

When a user enters certain details, like topic of interest, end-goal, preferred method of learning etc., a prompt gets generated and fed to an AI model. The AI model, based on this information produces a structured Roadmap, split into Tracks and further Tasks. The user completes tasks and so completes a track and then finally the roadmap. Every task has a deliverable the user needs to upload, the AI model analyses that to judge your level of knowledge.

How we built it

I used React components to build the major dashboard and task details page. The Login/Authentication was managed by Supabase. The database schema was also written into schema. Based on inputted information, we built a prompt that we fed to AI, that returned a structured JSON doc that was used to fill the rest of the database. The frontend interacts with columns in this database to display information in the dashboard

Challenges we ran into

So, so, so many. This is my first time developing a website, and I was floundering. I had no idea so much went behind authentication and log in and maintaining so many users, retention and so on. It was way more complicated than I thought it would be.

Accomplishments that we're proud of

I'm proud of whatever I've done. I got to know of this competition only last week, and all that i've been able to produce in this short week makes me extremely happy. All the 3 am coffees pay-off when you see the tiniest piece of code run with no errors on the first try. And I have tried to use as less AI as possible to make this a true learning experience for me.

What we learned

I learned debugging, database deign, defining goals, breaking a large goal into smaller steps. Technically, I learnt HTML, Java, CSS, using API keys, REACT, FIGMA

What's next for Atlas AI

So much, this is just the bare minimum, we can have personalized AI chats, weekly evaluation, monthly evaluations, a separate corner for research paper analysis and reading, summary of feedback on deliverables and so much more

Built With

• css
• figma
• html
• javascript
• openai
• openrouteservice
• parse
• react
• supabase
• vscode