Dream Engine

The global job market more competitive than ever, whilst job candidates continue to develop their skills

Comment

Inspiration

The global job market is more competitive than ever! Whilst candidates continue to develop their skills, many of these do not show on their CVs or get lost in conversation during interviews.

What it does

The Dream Engine extracts structured, valuable information from recorded job interviews (transcripts). The platform harnesses state of the art NLP along with our own algorithms and heuristics to give HR the insight they need, so that you can land the job you’ve been longing for!

How we built it

The engine implements the Google Cloud Natural Language API. and key features include

  • Element sentiment analysis to identify a candidate’s skills (e.g. A candidate might be proficient in "databases", "SQL" or "mySQL").
  • Sentiment analysis and our own domain specific heuristics to quantify a candidate’s skill level. (e.g. A candidate may have "years of experience" in one area whilst only "a little" experience in another).
  • Analysis of dependency trees to identify negation (e.g. A topic might come up in conversation although the candidate is not skilled at it).

Challenges we ran into

Dealing with multiple edge cases and understanding the complexity of real time natural language processing.

Accomplishments that we're proud of

We were able to dedicate our efforts to the challenges of NLP applied in a context with few data sets available, where standard machine learning can not be used. And not killing each other!

What we learned

NLP techniques and a lot about the semantics of language.

What's next for Dream Engine

Only time will tell, but Go hard, or home! - We went hard, now we're going home to Sweden.

Built With

Share this project:

Updates