An Interview with a Grab Inventor: Victor Lang

At Grab, our deep understanding of Southeast Asia has helped us evolve into the region’s leading superapp, designed to empower everyday entrepreneurs while continually innovating to provide our users with delightful experiences. In our latest series 'An Interview with a Grab Inventor', we shine a light on our Grabbers pursuing innovation driven by their passion. To foster the culture of innovation, inventors at Grab are also rewarded for their inventions and efforts contributing to our growth and expanding patent portfolio.

In the first chapter, we sat down with Victor Liang, one of Grab’s most prolific inventors, to deep dive into his work, the inspiration behind his inventions and their impact.

Victor is Head of Data Science, Geo Services. With eight years of experience in the data science industry after obtaining his PhD from The Hong Kong Polytechnic University, he's been contributing to Grab's innovation for the past seven years. With 14 patent applications filed at Grab, Victor stands out as a top inventor.

What do you think are the most valuable skills for an inventor?

I believe that great innovations often come from continuous improvements and iterations, and not sudden ground-breaking moments. The most valuable skills an inventor can have, in my opinion, are optimism and the attitude to keep on trying. Just continue to work on it without hesitation, even if you think that it is a very small increment, because you never know how and where your iterations will lead you.

Of all the inventions you have developed in Grab over the years, which invention are you proudest of, and why?

The invention I’m proudest of is the “Method and System for Gathering Image Training Data for a Machine Learning Model”.  This is a very innovative product feature on KartaCam, our in-house camera device and software for capturing street view images for GrabMaps. GrabMaps was created to address Grab’s need for a more hyperlocal solution (e.g., back alleys and narrow side streets, which are very common in Southeast Asian cities) to power its services. 

Each image we capture has a cost, so ensuring quality is crucial. The mechanism uses computer vision to spot poor-quality images (blurry, under-exposure, over-exposure) in real-time, triggering a retake automatically.

Today, GrabMaps is self-sufficient and provides location-based intelligence and services right from Base Map Data to Map Making tools & SaaS and Service APIs to all Grab verticals in all the eight countries it operates in.

What was the motivation behind this invention, and how did it come about?

We wanted to make every mile our driver-partners covered count by identifying and replacing poor-quality images in real-time. That’s why the innovation here is to leverage the AI capabilities on KartaCam to improve mapping efficiency. 

How do you get the spark for your ideas and inventions?

I usually try to approach problems from different angles. While it may be easy to have an intuitive solution, I challenge myself to figure out a solution B. This thought process helps me to find those sparks of innovation.

What is your advice to fellow inventors?

Having an idea is easy but having a good idea is hard. I usually challenge myself by pushing the problem to an extreme condition. If there is a decent solution that addresses the problem well, that could be a novel solution. This thought process helps me refine my ideas continuously.

Want to know the latest happenings at Grab? Join our Talent Community.