Through our own personal use of job search and professional networking sites, the suggested positions and connections seemed remarkably unrelated to our selected preferences and stated areas of interest. We wanted to help especially the student population, who has less experience navigating the job market at large, with finding companies that they would be excited to go to work at every day!

What it does

Oliver provides a more robust user-assessment process and utilizes social media content in tandem with machine learning techniques to find companies and available jobs highly personalized to the user.

How we built it

Using a combination of APIs, web scraping, AI/machine learning (built on Apple's Core ML/Create ML platform), and SwiftUI, we developed an iOS app for job-seekers to provide information and pull content from services to complete a profile that assess which companies have an environment that is best suited for their interests and skill set. All of the data pulled from various services that we did integrate with stays completely on device, and the actual ML occurs on the device as well—making this a platform trustworthy with so many external data sources.

Challenges we ran into

Several of the APIs we used for data collection required developer credentials that would take 1-2 weeks to request approval for. In light of this, we had to utilize web scraping for data which slowed down the development process and forced us to focus on particular aspects of the data integration that prohibited us from expanding the interface of the system as we originally intended.

Accomplishments that we're proud of

We were able to work around complications as they arose and stayed organized and flexible throughout our development process. We produced a working and very polished iOS application to demo our robust AI-based system. As we shifted focus away from getting as many integrations as possible, and more towards a thorough and well-performing recommendation algorithm, we saw a huge jump in quality in the recommendations we were serving. At the end of the day, we're proud of the polished user interface (which is somewhat of a luxury to have time to develop within the time constraints), the algorithm itself, and the ways we were able to split up tasks between the two of us. We're very happy with how much we accomplished as just a two person team!

What we learned

A great deal of insight was gained into the nuances of integrating multiple APIs from numerous web-based sources and utilizing machine learning techniques to draw connections between numerous data points and integrate solutions into our application.

What's next for Oliver - AI Job Search

We have many more plans for Oliver that, due to the nature and time constrains of the Hackathon, we were not able to implement. Further integration of various social media sites and content providers is certainly possible to amass a more robust database for even more personalized and salient results for the user. A partner app would be designed as well for use by companies utilizing data aggregates collected by job-seekers to give employers a more clear idea of what traits, characteristics, and environments employees are looking for in a workplace.

Built With

Share this project: