-
Abhyas dashboard with sentiment analysis and file recordings
-
Job Analyzer
-
Representation of the flow: Set job description, Track Progress, Read Curated Articles
-
Evolution of Abhyas over MHacks
Inspiration
Telephonic interviews are the first step in the recruitment process for any role - design, engineering, finance or any other domain. These interviews are not very technical in nature but rather focus on behavioral questions and gauge the overall acumen of an applicant in the field.
"Abhyas" or practice in Hindi, is the one way you can excel in this process and was our inspiration for this hack. We set out with the question, "Can we make this fairly straightforward practice smarter?".
What it does
Abhyas helps you practice smarter and gauge your progress over time. With Abhyas, you first pick a job role you are preparing for and enter the job description in the system. The system runs an entity extraction algorithm on this job description and builds a set of key skills that you require for this job.
You are now ready to practice! Dial in on the given number anytime you want to practice and Abhyas will give you a set of questions that you answer in real time. Following the interview, Abhyas gives you analytics on your answers based on the overall sentiment and also, how close or far your answers were from the originally uploaded job description by running a semantic distance algorithm. You can also listen to any of your recordings to see how you have improved over time!
How we built it
We build our telephonic interviewing platform with the Twilio API which enabled us to create a smart IVR that fetches new questions for the interviewer to practice and also to record and transcribe their answers. The user data is stored on Firebase, which we query in our web application to run the sentiment analysis and semantic distance algorithms (both implemented using Dandelion API).
Challenges we ran into
In terms of user experience, we had to decide on the flows of the entire process. Originally, we had planned to only create a system that lets users listen to their recordings but after doing some quick user research we realized that people wanted some sort of feedback. Determining how to give feedback to our users was a big challenge.
Accomplishments that we're proud of
Building a platform, designed with its end users in mind. Both of us learned a ton of new skills and used this opportunity to create something with meaning.
What we learned
We learned a lot during the making of this project. We got to explore voice user interface design through Twilio along with using Firebase and Dandelion API to power our ideas in the backend.
We had a wonderful time sharing and collaborating as a team. We had a lot of great ideas, which we saved over the course of the hackathon.
What's next for Abhyas
We would like to make this a hybrid application which can work across devices. Our future goal would be to allow people to call and get more controls to fine-tune interviews right away without relying on another interface, though, there will always be an option to use web/mobile interface.
Log in or sign up for Devpost to join the conversation.