Inspiration

A large number of people suffer the absence of a friend who can listen to their feelings without judgement. Depression rates particularly increased during the pandemic and there has been a rising trend ever since.

About 7 out of every hundred men and 1 out of every hundred women who have been diagnosed with depression in their lifetime will go on to complete suicide. The risk of suicide in people with major depression is about 20 times that of the general population.

In our team, we recognized the need for a private supportive companion that is both accessible and user friendly.

What it does

Nemesis is an intelligent daily diary application designed to be a private, supportive companion on your mental wellness journey. It goes beyond simple note-taking by using advanced AI to analyze your thoughts and feelings, provide actionable insights, and bridge the communication gap with your healthcare provider, all under your complete control. Noesis is your personal guide to become a happier being and make others happy too! It helps you to focus on maintaining good health and well being and also helps you to become a better component of a growing society.

How we built it

We started this project at 8:00 pm IST. We wanted to build something in accordance with this year's theme - Equality and social good. Being environmentalists and having done social service before, we were determined to put our creative efforts together to build something amazing that will improve the mental health of the society at large. We used trained AI models from Google Gemma. PyTorch was used for the NLP integrated elements and Tailwind CSS for the formatting.

Technical Implementation Summary

We developed a mental health analytics dashboard with the following stack:

Frontend:

  • React + TypeScript (Vite build system)
  • Tailwind CSS for component styling
  • Chart.js for data visualizations
  • Client-side state management (React hooks)

AI Components:

  • Google Gemma API integration (sentiment analysis)
  • PyTorch NLP pipeline (text processing)
  • Mock inference endpoints (local JSON responses)

Current Limitations:

  1. Production API Access:

    • Requires to be paid Gemini API subscription (unavailable to us)
    • Live model inference disabled
    • While testing, it could be used by devices connected to the same local network only
    • Therefore, we've also attached below a Google AI Studio link to test and utilize this application to its utmost capacity
  2. Local Execution Constraints:

    • Runs on localhost only (no cloud deployment)
    • Uses static sample datasets
    • Limited to pre-trained model snapshots

Functional Components:

  • Complete UI/UX implementation
  • Interactive data visualizations
  • Client-side analytics processing
  • Sample data workflow

Challenges we ran into

We were beginners at this. It was our first time in a hackathon. So, the pressure of having to finish something on time, fighting the sleep, and still trying to make it as memorable a project as possible was a challenge. While doing the editing and coding parts itself, the variables got jumbled up, which ended up wasting about 1 hour 35 minutes even after everything was done. All of us had only beginner level skills in coding. I think the strong determination to have fun and create something incredible drove us and helped us avoid any serious challenges.

Accomplishments that we're proud of

We are particularly proud of the diary feature and how the overall color themes and outlook turned out. We were able to directly go from brainstorming ideas to creating them in real life. When we look back, we are very proud of our teamwork and cooperative nature in which we approached this project.

What we learned

We learnt a great deal about computational linguistics. If it weren't for this hackathon, we would never have been able to approach AI from the creator's perspective. We learnt a lot more from this perspective than the user one. We also learnt that implementing ideas is not easy. But, of course, with effort and determination we can achieve anything! When we had the final project, we were really happy with the result. So, we also learnt that real happiness comes from putting your full effort into a job and seeing it done.

What's next for Noesis

We want Noesis to become a memorable collection of a person's daily life. So, we will definitely create more interactive elements and options to include your pictures. We will also include a video option to talk with Noesis, and it will give feedback and talk with the user based on their facial expressions and lines. We are also planning to extend this app so that a group of friends or workmates can use this to create healthy conversations and fast progress in jobs.

Built With

Share this project:

Updates