Inspiration
Our inspiration started with Chat-GPT and our astonishment at its capabilities, which convinced us of the role of AI natural language processing in the next generation of the internet. Paired with our memories from high school of the intensive research that went into researching universities, we came up with the idea of using AI as an assistant for helping find relevant university courses. This later morphed into the idea we have now of generating job recommendations from a CV and a few prompts.
What it does
ResumeAI takes as input the user's CV in the form of a PDF file in addition to a few prompts to further aid in searching for a relevant job - The field, specialization, commute distance, and general area. The inputs are first taken to Selenium, which scrapes a range of broadly relevant jobs from indeed.com. These are then fed into GPT-3 using a customized prompt to determine 5 jobs which best fit the user's given criteria.
How we built it
Our project is built in its entirety using python libraries. We used Selenium to scrape a range of jobs from indeed.com and OpenAI's GPT-3 API to determine which are most relevant.
Challenges we ran into
One time-sinking challenge we faced was determining what front-end to use for our project. Before settling on Streamlit, we experimented with interfaces built in SwiftUI and discord.js, however we ran into challenges with integrating our python APIs with the interface given the time constraint.
What's next for ResumeAI
We still have plenty of ideas of how to continue improving ResumeAI: The biggest planned improvement would be to provide more detail to the user on how GPT-3 is determining why it chose the top 5 job offers. This would involve fine-tuning the model to provide more context on its decisions and a weighing of its choices.
Log in or sign up for Devpost to join the conversation.