The Game is On.


It’s a well-known truth that interview preparation is often a shot in the dark. You prepare for certain lines of questioning before the interview and leave flummoxed by the time you’re done. We set out to resolve this enigma by quantitatively analyzing what companies are really looking for when you walk through that door.

What it does

The app asks the user a number of questions that the specific company often has asked interviewees in the past. Users then record audio responses to those questions. The responses, along with scraped data from the company’s website, are sent to Watson for Tone Analysis which delivers linguistic, social, and emotional analysis on data processed. Through data visualization, the user is able to see how his answers stack up against what the company is looking for on a number of characteristics using spider charts. The user then has a quantitative understanding of how he needs to change his answers to succeed.

How we built it

We used Swift to build our iOS application, as well as to interact with Watson on Bluemix and the iOS Charts API. The application uses iOS’s native speech recognition to convert the speech input to text. We then used Watson’s iOS SDK to interact with the Bluemix API to send requests to the system. After parsing the JSON output from Watson, we displayed that data on spider charts.

Challenges we ran into

Interacting with Watson was a major challenge largely due to the fact that we had never used any kind of similar platform before. However, once we started understanding the processes that were occurring under the hood, we gained a good grasp on the system. In addition, we initially struggled to display our data in an original manner that captivated the user’s attention while communicating as much information as possible. However, after modifying a number of different chart APIs for iOS, we were able to arrive at a data visualization which we felt achieved our objectives.

Accomplishments that we're proud of

We're proud to have developed an application that utilizes the power of Watson on Bluemix towards a problem that nearly all professionals face in their career. We are happy to have attempted to solve a problem as relevant as this however challenging it was as it has the potential for wide usability in the community as a whole. It was also enriching to increase our development skills by attempting a project outside of our original capabilities.

What we learned

  • Advanced Features of Swift
  • Watson on Bluemix
  • Customizing APIs for Personal Use

What's next for Sherlock

As a further development, we would love to crowd-source answers from successful applicants to specific companies and then use machine learning so that Sherlock also has a native understanding of what kind of things a company looks for in interviews as opposed to data scraping being the only source of data for the app. We are keen on expanding on Sherlock beyond this event so that it can become a powerful tool for the community.

Share this project: