More companies started using video interviews to select ideal candidates for their roles. Our project,Interview-101, is tailored for those who are not confident with their interview techniques. As computing and engineering students, we find the video and face-to-face interviews in a job application the most challenging and we feel that there still lacks an online, user-friendly platform for us to practise on video interviews. Hence, we have come up with the idea of using Azure features to analyse interview performance, with key metrics specified, so to empower users to excel in a video interview.

What it does

Our web platform allows users to practise individual and group video interviews. When the user enters the portal, he/she will choose the interview type , and will be given a timed mocked question drawn from our SQL database which consists of real-life interview questions . The performance of the interviewee will be analysed; results and feedbacks will be visualised on a dashboard, displaying key metrics such as the talking speed,sentiments, cliche words, and clarity, generated using Azure Cognitive services and Machine Learning. Another distinguishing feature of our portal is group interviews. Real-time communication technology is leveraged, allowing multiple users to participate in the interview.

How we built it

The web portal consists of a frontend, which was developed using React (material UI), a server built using NodeJS and Express, and a Python backend integrated with custom algorithms and Azure Services. We made use of the cognitive services such as the Microsoft Video Indexer to analyse texts and languages. Along with our custom algorithms (such as linear regression models), we determine influential factors such as the confidence level ,sentiment level, time of pauses and stutters, and provide constant feedback to our users. We analysed the video transcript to filter keyword, and compare these words with the integrated SQL database- through collecting and storing key responses from our users, we intended to find out the most cliche answers for each interview question, hence to provide an insight into how candidates could stand out in an interview. Web sockets and web RTC technologies enables us to build a real-time interactive platform for multiple interviewees, distinguishing us from the current market . The project was developed in Microsoft Visual Studio Code and was deployed using Azure VM Cloud Service.

Challenges we ran into

Timing, workload, and health condition were our main challenges. Our team had set a very ambitious goal- integrating front-end, back-end and additional features, resulting in a sleepless night for all of us. At the beginning, we struggled to find relevant datasets to train our models and we wasted a lot time ideating. Then we faced with the challenge of integrating no-sql cosmoDB with our nodeJS application; despite seeking helps from the sponsors, we were still unable to integrate it.

Accomplishments that we're proud of

Our team was able to develop a functional platform , leveraging various Microsoft Services, within the time limit. We were particularly proud as we had great work division and everyone was able to contribute. We were able to identify the limits of current solution (e.g. not very user friendly as users often have to upload videos to the portal ) and built a more advance and more customised product. We have learnt a lot and were able to polish our skill in programming.

What we've learned

We had a great, well organized hackathon experience with the supports of very helpful and engaging sponsors. We got to make practical uses of cutting edge technologies to alleviate real-life problems. We have learnt how to work efficiently, how to ideate ideas so that we can target real-life issue and to empower others.

What's next for Interview 101

As we could see a huge business potential in this idea, we would like to expand our database to more questions (such as from previous video interviewees) as well as answers from successful video interviewees, and provide a platform of sharing information between job-seekers. We have also started building on doing online group interview and analyse group overall performance, with Azure’s human interaction feature.

Share this project: