Inspiration

We took inspiration from MyCareersFuture. We wished to make it more accessible, while simultaneously conveying the information more concisely. While we may have taken inspiration of their features, we used a minimalistic UI in contrast to their colorful UI to avoid visual clutter.

What it does and how it is different

  • We allow our users to upload their resume to be converted into a detailed form, to expedite the process of user-job pairing.
  • Keyword-extraction of both job description and the users portfolio has been implemented to increase the accuracy of matches.
  • We used our custom-prompted LLM to churn out a highly summarised and efficient job posting from available postings on third-party services. Through this process we are able to cut down the time of the application process by reducing a large amount of filler words.
  • Real-time filtering of job results through custom parameters provided by the users.
  • We use ATS to judge our users' portfolio and to match them with their preferences and with the employers that desire a similar skillset.
  • We use AI to suggest a timeline to upskill the applicants resume in the meantime.
  • AI insights into your suitability for the available roles.

How we built it

We built the frontend with Next.js and Tailwind CSS, managed data with a Neon (Postgres) database and Prisma, and integrated Llama for our AI logic. We worked hard, collaborated, brainstormed, and guided each other in our own specialties in web development. We took aid from our available resources not limited to LLMs, StackOverflow, and the language/service documentations themselves.

Challenges we ran into

API key and rate limits, sleep deprivation, gemini going insane after failing to provide a csv file. Towards the end of the night we faced a severe lack of energy in all forms, mental and physical. This led to a heavily reduced rate of production. Over the course of our project we also ocassionally faced issues with our choice of platform, VS Code LiveShare due to network and hardware issues.

Accomplishments that we're proud of

We are proud of being able to create and deploy a working website in less than 30 hours. As first-year students lacking experience, such an achievement was exhilarating. We were able to develop both the backend and the frontend to achieve an all rounded platform.

What we learned

API integration and interaction with both the backend and frontend. Learning to handle rate limiting among various LLMs, prompt engineering to achieve desired results from our LLM API. We also gained valuable experience with both the backend and frontend side, getting more experience with various web development toolkits.

What's next for Team 73_FindJobAlready

Wider API integration with job searching websites with more accurate upskilling roadmaps with more interactive elements. We also hope to make our features more efficient so they run smoother and faster.

Built With

Share this project:

Updates