Running out of resumes at a career fair was frustrating. Especially when it happened nearly every time.

We wanted to make it easier for students and recruiters to trade contact information / resumes and keep the entire process streamlined.

We first decided to pull information straight from LinkedIn by comparing a photo of a person with profile pictures on LinkedIn. This would obviously take forever given the number of people on LinkedIn. To optimize this, we went back to looking at the way career fairs were done.

Registering for a career fair. Yes, that was it! Whenever a student registered for a career fair, we could ask them for their LinkedIn profile and resume and store that on a database. That way, when someone walked up to a stall at a career fair, we could match their picture with everyone who registered.

Implementing that was pretty straightforward and we built a proof of concept app that could show some information about people based on their picture;

What it does

Snap a picture of a person's face, and our app will find the person's Linkedin profile, name, and summary.

How I built it

We wrote an Android App in Java using Microsoft's Oxford API and Linkedin's API together. We used Linkedin's API to pull images and other information from user profile pages and Microsoft's Oxford API to compare pictures to match the right information to the right user.

Challenges I ran into

  • Microsoft's Oxford Java SDK did not have proper documentation. Instead, we had to go into the source code to figure out what each method did and how to make the calls properly.
  • LinkedIn's API was also a huge challenge to work with because they required strict security measures to be fulfilled before allowing access to any of the data.
  • Finally, we ran into a lot of bugs when we used multithreading to make our UX seamless. Every API call runs in it's own AsyncTask.

Accomplishments that I'm proud of

  • Being able to make a fairly accurate facial recognition centered app.
  • Making our identification algorithms efficient enough that they run fast enough to make a viable app.
  • Placing second place in MLH Stackathon Stackers

What I learned

  • Facial recognition is cool
  • Not all APIs are well documented
  • Not all heroes wear capes
  • Nidhogg is a fun game

What's next for FacedIn

Besides expanding the scope of FacedIn as well as improving search times and efficiency, we want to add the feature to save records of students so that you can go over them post career fair.

We also want to be able to integrate other services like Twitter and Facebook and get a student's resume. This would require a fairly advanced web crawler and may not be practical but it's worth a shot.

Really, we want to be able to fetch your online identity based off of a picture. Yes, this is NSA level scary.

Share this project: