In these trialling times of COVID-19 where millions of people are desperate for jobs, the last thing a candidate wants is to not know where they're going wrong. Interviews have switched to be virtual and candidates feel the lack of a human sense while talking to a camera. Often times, certain rounds are pre-recorded on platforms like HireVue and candidates tend to lack eye contact or expressions. These platforms compute such factors but candidates have no resources to prepare for these aspects. PrepHR aims to address this issue and give candidates a platform to practice for virtual interviews. PrepHR also serves another purpose - to help practice for behavioral/HR interviews with Mock questions asked in real interviews.
What it does 💯
- Enables practice with automated behavioral interviews and displays a performance score rated on different parameters such as confidence, posture, eye contact, etc.
- Allows practicing infinite interviews and tracking your progress on your personal dashboard.
- Has a rich question bank from several categories to nail your next HR interview
How we built it 🐍
With ReactJS, Bootstrap, CanvaJS, Django, MongoDB, and a lot of sincere efforts!
Challenges we ran into 🌪☀️
- One teammate had never worked with ReactJS before.
- Our teammates were in almost opposite time zones and hence time coordination for meetings and task checking was difficult.
- Finding the right API and trained model for the purpose of scoring was difficult.
Accomplishments that we're proud of 🌈
- It is a working solution with key features implemented.
- The machine learning models to score a video are working.
What we learned 📕🎓
- We did significant work on Figma, which was new to us (https://www.figma.com/file/2UivqSl9HnkscnGExh5chx/PrepHR?node-id=158%3A2)
- Working on a ReactJS project at this scale was also a new experience for some of us
- CI/CD was fun and interesting
What's next for PrepHR ➡️⏭️
We have the following 3 goals for PrepHR as future aspects
- Live Mock Interviews with a random pairing of online users or a real person reviewing and sending feedback
- “Correct” answers based on keywords
- Avatar using StyleGAN2