CareerAdvisorAI

Inspiration

I am a student that likes CS and logical problem solving, but has a poor understanding of the job market and what the workflow of industry specializations actually entail. I built this tool to be a tool I wish I had access to when beginning to think about what I want to study.

What it does

provides specialized, personalized, and in-depth career advice as if from an industry professional, based on information the user provides through conversation with the AI agent.

How we built it

AI studio agents were used to create a skeleton, then the gemini api was used in a python backend framework.

Challenges we ran into

the main challenge was maintaining consistency in the style and quality of responses when first attempting to replace the temporary job listing data file with actual data gathered from webscraping.

Accomplishments that we're proud of

I was able to string together a working proof of concept with no real development experience prior to this event

What we learned

Unrestricted vibecoding can often lead to errors that are difficult to debug since the organizational structure of the program can be difficult to follow for a human reader. This is why I switched to using detailed organizational instructions to micromanage the logical structure of the program, so if there was an error when connecting the backend, it could more easily be recognized and debugged.

What's next for CareerAdvisorAI

I want to flesh out the app over time into something that provides a varied and seamless user experience. this includes optimizing the prompts to provide more reliable/relevant responses, adding other tools that aid in applications, cover letters, and resume optimization, and continuously scanning for new job opportunities and providing push notifications for ones that may be especially relevant.

Built With

• python