Broast

"Don't play with Broast!". An angry Broast, when you click nearby him.
"It's so boring when there's nothing to roast". Sad Broast, give him some tea to spill!
"Finally, something!". Happy Broast after submitting the form.

Inspiration

Our application grew out of the curiosity to take a fun spin on the oftentimes long and arduous task of job hunting. This came about in the form of an application which will roast the applicant's resumé (in a nice enough way, we promise).

What it does

Broast will take in a submission of a user's resumé, after which our lovely Bear companion will begin to roast it, while providing advice to improve it. Additionally, the user will be guided to various job positions which best match the information given within their provided document (resumé, cover letter, Github page, or LinkedIn page.

How we built it

Broast uses a Python-based back end, built using the Flask framework, to handle routing requests, OpenAI API calls, and Perplexity integration. We used GPT-4o-mini as our model to generate our playful insults, glow-up tips, and compliments. Perplexity's API was used to receive pre-determined prompts to generate a list of recommended job titles, and related job postings based on the user's resume. The front end is built using the Angular framework to generate tabs and toolbars, and of course, our friendly and willing Bear participant!

Challenges we ran into

While we are happy with the results, several challenges would involve getting each aspect of our application to integrate as a cohesive whole. For instance, integrating Perplexity's API to work with the rest of the application proved to be a tricky task.

Accomplishments that we're proud of

At an overall level, we're proud of the overall scope of the final project given the rapid development schedule, and as second-year students learning several concepts regarding AI/ML concurrently with the development of the application. At an application level, we're proud of of the way we've built our application to bring together front and back ends, alongside Perplexity and OpenAI integration to enhance the user experience. Not to mention, we love our bear companion, he looks mean but he's quite nice deep down!

What we learned

Our experience has given us greater context as to what can be feasibly developed given a number of factors. Those major factors being skillsets; on an individual level and as a team overall, and time-constraints. On a purely technical level, we've learned how to use Perplexity and it's APIs within a project, and we've gained insight on several ways to integrate multiple parts of an application into one cohesive whole.

What's next for Broast

We were thinking along the lines of adding user account creation, along with security features to boot. Additionally, we would like to work on adjusting the formatting of the output, particularly regarding the links generated by Perplexity.