Elias Vafiadas, Head of Software Engineering and Technology, AirAsia Group talks about the usage of data and machine learning to arrive at groundbreaking results.
Vaishnavi J Desai | ETCIO | April 03, 2019, 07:00 IST
AirAsia recently launched biometric facial recognition technology at the Senai International Airport, Johor Bahru, Malaysia. The move comes in the wake of digitization effort by the low cost carrier. While digitization does bring in a lot of ease, it also generates data. When structured and put to good use, it can lead to insights that can help an enterprise arrive at decisions faster, save costs and achieve groundbreaking results.
Elias Vafiadas, Head of Software Engineering and Technology, AirAsia Group talks about the data as the new gold, devops and its usage, and the availability of skillset in the markets that the technology centers have been opened in.
A lot of data accumulation happens because of the technologies that you’re implementing, namely machine learning or facial recognition technology at the Senai International Airport. How do you use the data to make decision making faster in the company?
Between machine learning and facial recognition technology (which was implemented to streamline the boarding process and operation), I think part of the digital transformation of AirAsia as a whole, not specific to the software engineering aspects – using data for decision making is key. So over the last one or two years, the teams have done a great effort to consolidate the sources of data to make it easier for us to view the data and be able to understand what the data is telling us and therefore taking decisions. There are multiple aspects of that in the airline; like the route scheduling and the seasons around that, in flight aspects such as how many meals do the cart have, the historical trends, what is the most optimal operation, etc. So we’re looking to leverage the data as well as machine learning, because parsing through a lot of data manually is time consuming and at times not an accurate task. So with the use of technology and machine learning, we’re trying to create models to predict the behaviors of particular aspects. Also, the biggest problem that an airline faces is fuel consumption. So if we were able to use technology, machine learning and the data that is produced by all of the various sources (the aeroplane has about 200,000 sensors; like there is a whole lot of data of the aircraft, the operations, the website, all across the journey of the customer from booking to arriving to the destination that we’re all trying to use) we can use that to solve some of the problems and it would be really groundbreaking and cost saving.
How are you putting DevOps to use?
In the traditional sense, software teams would have a development team that will produce software and then they will have an infrastructure and servers team that they will take the bits, put them on the server, then monitor them, etc. So that model has worked well for years. But now the industry has shifted into a more continuous delivery, continuous integration model. And this is what we’re heading to. Our new website is leveraging on cloud technologies and micro services and things that help us accelerate the duration on those products. One of the components that we’re investing as we speak is the ability to ship software, immediately or as close as possible from one developer premise to bringing the operational aspects closer to the developers. And that’s why we can iterate much faster and our quality will not be compromised as we increase the speed.
In terms of app development, customers in India are also in tier 2 cities, tier 3 cities, where the connectivity is still low. How do you make the app light or tackle the challenge?
It’s not just India there are like many markets like that where connectivity is a challenge. And this is one of the reasons that one of our first initiatives in this digital transformation was to basically take a deep look at our booking engine and rebuild it using progressive web apps which is meant to solve the problem of connectivity. First off, a lot of the components get downloaded on the device in a shell, where, you know, the basic connectivity is there. So as you move from screen to screen, you can still interact with the application without having to wait, to load every single thing over and over again. So that’s live now, and we will continue to improve it. So this year, you will see many changes on the app, you’ve already seen a couple months back when we announced the facelift of the mobile application which has a new look and feel on the homepage and the flight search for iOS where it looks different. This leverages the technologies that we’re building for web which are lightweight, and we’re continuously looking to minimize the payload of our APIs, to make sure that the mobile app will be really easy to use, whether you’re in a normal or poor connectivity.
Do you think that there are sufficient skills in the market that you have opened the technology center?
We definitely opened the tech center in Bangalore because we believe there is a great talent. So there is definitely a great talent pool that we’re looking to pull from and build our tech presence in this region as well. Lots of people know AirAsia as an airline and so you will be seeing us more and more in university events, in conferences, meetups, all of the different mediums in all of the locations, so definitely Bangalore and Kuala Lumpur and Singapore to a certain extent. We’re taking initiatives to build a strong talent pipeline to maintain our growth and expansion.