Johan Data Scientist Grab - Pull out quote 

Johan Kok's interest in data science began in his undergraduate days as a Mechanical Engineering exchange student at the University of California, Berkeley. The experience gave him a taste of the world of computer science, and Johan was hooked on the fast-moving culture and the constant need to upgrade himself in face of the ever-changing technology. He was excited by the potential impact that he could have on society, with just a laptop and knowledge in coding.

Upon returning to Singapore, Johan began studying computer science in his free time. Now, as a Data Scientist with Grab and a PhD candidate at the National University of Singapore, Johan spends his days researching deep learning models used to forecast data warehouse resource demands with technologies like Tensorflow and Python, and also pre-processing data pipeline tools like Spark.

To learn more about his career in data science, we sat down for a call with Johan. 

How did you come into this role at Grab? What drew you to it?

I was applying for jobs after graduation and Grab was one of the companies that I applied to. I was intrigued by Grab’s selection process. There was the standard interview, and then there was the “build a pricing mechanism in 2 weeks” assignment. This was unlike any other interviews.

I really appreciated being assigned a mini-project, as it not only allowed me to dabble with technologies that became the focus of my career, but also gave me a clear sense of what being a data engineer entailed. In the end, I decided that my interest was exploring the field of big data and accepted the offer.

How do you think that interest has led you to your current role?

I was very much drawn to research during my undergraduate days. My first exposure to building and training machine learning models was during my internship at A* Star, where I worked on deep learning projects for cancer detection.

I was encouraged to pivot my focus towards big data at Grab. Being a data engineer at a company managing insane amounts of data gave me exposure to the techniques in building systems that can handle any amount of data without losing data points. It also showed me real-world challenges with big data systems, motivating me to dive deeper into the field to solve them.

Johan at Grab with colleagues

What is it like being a Data Scientist at Grab?

Most of the work we do here is self-driven, so I have a lot of free time to explore my area of interest.

Data science is a tough field to venture into. It's 10% fun and 90% hustle. You need to understand the key characteristics of the data that allow your model to excel, and build pipelines that clean and scrub data so that they are formatted to be compatible with the model that you developed. Be mentally prepared for this.

What are some interesting projects you are currently working on?

I’m working on a project which applies deep learning models to the prediction of cost metrics, such as runtime, for an incoming query without having to execute the query and profile the results. Current state-of-the-art models often fail to factor in the complexities of distributed query engines, and do not dynamically adapt to changing data characteristics, such as data size.

A model that works for any distributed query engine can be deployed to help with resource allocation strategies in data warehouses, which is especially useful for companies - like Grab - that power their applications on cloud-based providers.

Have you faced any challenges at work?

There are many challenges. Overcoming them was a matter of willingness to accept criticisms and improving myself, as well as not taking the easy way out and being thorough in my learning. I strongly believe that being good at a specific technology is a worthwhile investment. It may take time to be proficient at it, but the benefits will be reaped when you need to optimize or debug it for your application.

What are some qualities that would make a person a good fit to be a data scientist at Grab?

  • Be comfortable working with minimal guidance
  • Be very thorough in understanding the technology you work
  • Be willing to learn, unlearn, and relearn

I can’t stress the last point enough, as the computer science industry, and especially in top companies like Grab, things are perpetually changing. You have to change along with it too.

What’s next for your career? Any dream projects?

I do want to spearhead and become a core contributor to a major (Apache level) open-sourced project. I also have dreams of giving technical talks in conferences for software that I work with. 

Johan’s passion for computer science and big data has led him to a fulfilling career which allows him to research, discover and build innovative data systems that will help to solve real-world challenges. If you’re keen to embark on your own tech journey, here are some programmes to help you advance your career in this exciting industry:

Published on 21 August 2020