For the last 2 months, we have been working to bring a couple of small and big improvements on AirCTO. One of those steps was to bring more transparency in picking the experts to do interviews.

Pick experts to conduct interviews

Once the customer uploads resume of the candidates to our system, our team use to check the experts on our system to see who could interview the given candidate based on the skill set for a specific job. So, this process has been opaque for our customer and they didn’t have any direct say in choosing the experts who they want to conduct this interview with. We wanted to fix this in order to achieve two things:

  • Bring transparency and give this power to our customers
  • Make the process smooth with no manual overhead.

So, now with our new system - once the customer uploads resume of the candidates to our system, they get to the next step where they see a list of all the qualified experts who could do this interview. We only show relevant experts by picking all the skill set mentioned in the job profile and matching it with our experts skill set in our database backed by powerful Latent Semantic Analysis technique behind the scene (We still have some engineering challenges on getting the perfect match which we’re constantly working on to improve). This is how the expert selection page looks like:


Once clicked on schedule now, it will ask for your confirmation for the medium of the interview and if you want the coding test to be done.

We suggest our customers to pick 2 experts - Primary and Secondary. For any reasons, if the primary expert is not available then our system schedule it with the secondary expert. The customer has an option to “Favorite” the experts for a particular job. It helps them to quickly select these experts for further interviews. We have also got an option to view “recent” experts which they had chosen earlier to conduct interviews.

Automated Interview Scheduling

We have come a long way from manually recorded interview phone calling to completely automated interview calls, but interview scheduling is something which we did manually. Whenever a resume would be uploaded in our system, our team would call up the candidate to ask his availability and then check the availability of the expert. If both of their availability matched, we would set up the interview from our system.

It involved a lot of manual work! Well, we are a big fan of automation and after introducing automated calls, we decided to implement the same in our scheduling process. Automating the scheduling process has its own set of challenges, mainly because there were two different parties involved; the candidate and the expert. We had a lot of brainstorming sessions on the entire flow of the process that went through 2 iterations and then we came up with the best possible solution.

As soon as an expert is chosen, our system triggers email and SMS to the candidate in order to get his three availabilities based on the priority of the interview. After the candidate provides his availability, there would be another email sent to the expert suggesting him to choose any slots they are comfortable with that the candidate has provided.
If the expert confirms a slot, the interview would be scheduled automatically. Else, the expert could request for more slots from the candidate.

The idea was to get out of the way of interview scheduling process and that’s what we have tried to achieve through this new process. We finally have our scheduling mechanism fully automated. This feature cuts down any manual interventions from the company and us. Also, it offers more flexibility to the candidate and to the expert.

Recorded Video calling

We have already implemented automated phone calling a couple of months back, but we didn’t want to stop at that. Until now, we had been doing the video calls explicitly through the Google Hangout. But now we wanted this feature to be tightly coupled with our system where we would have more control over the features. At first, we decided to build our internal video calling platform but as we dug deeper, we realised that is a complex process when it comes to recording the video call. So, we kept the idea on hold for some time.

After a bit of research, we came across Tokbox and realised this is the platform we want to go with. We integrated the video calling functionality on top of ToxBox API. It was a great experience working with it. Since the API was well documented, the integration worked seamless for us. There were a couple of things we had to manage on our side like maintaining the presence and making sure if both candidate and the expert are online then only allow them to connect. Since recording both sides of the video and voice together was important for us. We faced issues where it recorded one party for 15 mins since the other party didn’t join the call. So, we quickly rolled-back and resolved this issue which works great now.

We have also released our own real-time code editor for conducting coding test during the interview process. Earlier, we were relying on Kobra/JSfiddle/Collabedit/etc for real-time coding tests. But now, we have got this working using FirePad on top of Firebase database which offers more control to sync tests within our system.

We will back with new updates on AirCTO very soon. :)