This is an example of the output for the ML standalone developed at CalHacks
This is an example for uploading a document.
As college students, we know that internships are extremely valuable as it provides us with the opportunity to take the theoretical knowledge we gain in class and apply it to the real world. However, the process of finding internships is slightly flawed. The career fair allows for quality in-person interactions between students and recruiters, but we have found that there are certain opportunities we can capitalize on. On the student end, we found that students often have to wait in hour long lines to meet with just a few recruiters on the off chance that they can fit the career fair into their busy schedule. On the other hand, recruiters have to pay exorbitant fees to attend career fairs, and it can be very expensive when you take into account registration fees, traveling, room and board, and so on.
What It Currently Does
NXTPitch allows for students to upload short elevator pitches similar to the ones they use at career fairs when meeting recruiters. Students can then upload supplemental videos, links to important social media such as LinkedIn, Github, etc, and have the opportunity to set 4 skills. Recruiters can then look through a feed of these videos; if they find an individual who fits their skill-set, they can look at the rest of their profile, chat with them and even schedule an interview. Furthermore, recruiters can also search for certain skills that they are looking for in a candidate; the system will then look for individuals that have those skills listed in their profile and display them to the recruiter.
What We Built
For Cal Hacks 2017 we decided to do a continuation project, and build an additional feature for our application. We created an AI backend, that enables students to receive automated feedback regarding the effectiveness of their videos and resumes/cover letters. Our AI backend analyzes the text and delivery of your content, and provides you with personalized analytics regarding how effective your content was. We break effectiveness down into three categories: emotional analysis, highlighted content, and professionalism and provide quantitative metrics for each of these categories using state of the art machine learning technology. We want to enable students not only to find jobs faster but also to find better jobs.
How We Built It
We built the application using iOS Swift for the user interface, Firebase for the main database, and a Python Webserver backed by Google Cloud for the backend. We also utilized various Machine Learning API's such IBM Watson and Microsoft Azure, for facial and textual analysis. The AI backend relies on many different Machine Learning API's, Python packages, and custom functionality to enable audio extraction from video, textual extraction from PDF's, and analysis of the content. We believe that our main feature is our ability to tightly integrate all these different services into a single application that provides immediate value to our end user.
Challenges We Ran Into
Since this was the first project we developed in iOS Swift, it provided us with various challenges as we had to get accustomed to language. In addition, implementation of the Python Webserver proved to be a challenge as it was our first time porting Python code to a server backend. Working with all the different API's was also a challenge, and ensuring the robustness of our application required many layers of error checking.
Accomplishments That We Are Proud Of
We are proud of the Machine Learning engine that was built over the Cal Hacks weekend. It is our most sophisticated feature and we managed to get an MVP of it within 36 hours. It already has impressive functionality, and should help differentiate our product from anything else out there.
What We Learned
We gained experience with a wide range of technologies such as iOS Swift, Firebase, Google Cloud, Azure, Flask, and so on. Additionally we gained experience in working with non-trivial API's in a demanding environment. Finally we also learned how to properly configure a backend infrastructure.
What's next for NXTPitch
NXTPitch is currently running small beta testing at UCLA with a dozen students and a few campus recruiters. If you're interested, feel free to contact us!