Connect

A job portal where recruiters/applicants can can connect using our meeting services. Get detailed analytics of how your conversation went with our services.

Comment

Inspiration

Giving/Taking interviews is a hectic process. Every company uses different software for meeting. After that comes the closing screen, which makes us anxious as to how the interview went. Here comes Connect, the unified job portal, where you can conduct meetings from our services. Afterwards, you can also get insights and analytics about how your interview went, from our services.

What it does

It's a unified job portal. Recruiters/applicants can use the meeting service we provide as the one stop meeting software provider. The meeting service is no less than other software like MS Teams, Google Meet, Zoom, etc. It provides mute/unmute audio, open/close camera, screenshare, customized background and many more things. After giving the interview, you can get analytics and insights of your conversation. Just upload the audio/video file, and our services will provide you with the insights.

How we built it

The backend was built using Django. The meeting service was built using the Twilio API. The conversation analytics service was built using Symbl.ai

Challenges we ran into

We ran into a lot of challenges:

  1. Integrating React+Django. Django serving as the backend and react frontend, we had to integrate them so that they communicate with each other via API calls. This wasn't an easy task as it replaced the original Django's authentication system. We used JWT to authenticate the users. Also, we had to set up CORS configuration as we were requesting resources from another domain.
  2. Working with AWS S3 buckets for file storages, boto3, accessing objects. We had never used AWS S3 buckets before. We even signed up for the first time to AWS during the hackathon only. We had to run through various documentations, threads to understand it. Some of our code wasn't working and after digging deep and spending 5+ hours, to our surprise we found in one of the forums that the documentation had given a wrong code snippet about Python AWS SDK. We later resolved the issue by following that forum. Also, accessing AWS buckets objects from our project was a hectic task as this was our first time. Sometimes it should not authorized and sometimes it should invalid signature after we resolved the first issue. This was also later resolved by digging deep into forums, docs, threads and changing the signature method.
  3. Working with Twilio video API, integrating it with our Django backend server. Using it for the first time, we had to go through various documentations of the Twilio video API. Many introductory and hackathon tutorial video are there by Twilio, but they are generally about text message and phone calls. We had thought to do something different which was indeed difficult. Understanding how it actually works itself took us 24hrs+. By this time when were able to get an idea about its API, it seemed to us pretty interesting, but we realized we were short on time, so we didn't add much features and will add in future updates.
  4. Working with Symbl.ai, generating tokens, accessing and using its API. After struggling to get a file url as mentioned above (point no.2), we had to send it to symbl.ai to use its services like Speech To Text, etc. This was a 3 step process, generating access token, getting a conversation ID, and then sending both these credentials to get the transcripts. However our API request returned success message with a 200 OK response, but to our surprise it was empty! We tried using some other service like "Topics", but even it didn't work. Then we dug deep into different docs, and slack channel, changed the entire structure of our project, they way it worked, and we did make progress, but it was too late (as this was just the last step of hackathon), and we decided to add it in the future to our project (after our hackathon)

Accomplishments that we're proud of

  1. We were able to learn a lot of things.
  2. Was able to make a video meeting service and a service which gives us insights of that converstaions

What we learned

  1. Integrating React+Django
  2. Working with AWS S3 buckets for file storages, boto3, accessing objects.
  3. Working with Twilio video API, integrating it with our Django backend server.
  4. Working with Symbl.ai, generating tokens, accessing and using its API

What's next for Connect

  1. Adding more advanced insights for conversations using Symbl.ai
  2. Adding more features to video like chat
  3. Improving the UI/UX, Design
  4. Adding a section where recruiters can list open positions and applicants can apply for them
Share this project:

Updates