RezoomAI

My Logo

Inspiration

Many CS students around the world have to apply for hundreds of jobs just to land one job interview. One common tip job seekers are told is to tailor their resume for each application based on the job's description. However, rewriting resumes for every job you apply to is a time-consuming task, especially if you're applying to hundreds of job postings. RezoomAI aims to fix this issue by automating the process of resume customization using AI.

What it does

Users enter their name, contact information, work experience, education, skills, and projects into a form. They also enter a job description. The LLM generates a brand new resume that uses keywords from the job description while optimally formatting the user's experience. Information input page

How we built it

We decided to utilize AWS Bedrock's Meta Llama 3.2 3B LLM to generate optimized resume bullet points. We used React/Bootstrap to create a user-facing interface to collect information for the resume. This information is sent to a database to store their account and user data. This data is then sent to the LLM to write strong resume bullet points. These bullet points then go to a backend where they are formatted into a 1-page document and sent to the user.

Challenges we ran into

We struggled to find the right framework and structure for the database. We tried Flask, Conda, and MongoDB, but these options were not optimal and quite redundant. We also struggled to make them work for user authentication. We also had to rewrite the prompt multiple times, because the LLMs we tested it with didn't give the desired output. Making the user-facing interface dynamically changing and sufficiently customizable was a struggle. Complications with our backend progress set us back greatly, so we compromised with fewer input fields, having to make a less sophisticated prompt and product.

Accomplishments that we're proud of

For most of the team, this was the first hackathon we had a finished product to submit for judging. We were also learning most of the technologies and frameworks from scratch, so creating a functioning application at the end was rewarding. Adding an LLM to the interactions between frontend and backend was also a new experience.

What we learned

All of the frameworks and technologies used for this project. Breaking down abstract ideas into actionable tasks. How much caffeine a person can ingest in 24 hours without dying.

What's next for RezoomAI

Another important step of the job application process is reaching out to recruiters and full-time employees. We want to figure out a way to search LinkedIn and other professional networking websites to find people at the companies we are applying to and help write a message to reach out to them. We also need a way to increase cybersecurity.

Logo