My Big Hairy Audacious Goal with AWS DeepRacer (Part 1 of 3)

This is a story of how my quest to acquire knowledge in Machine Learning, led me to AWS DeepRacer, which in turn led me down an unexpected journey of Entrepreneurship over the next 9 months.

Truth be told - the word "Entrepreneurship" is not usually associated with someone who has spent the last 15 years of his career in a bank. But at DBS Bank, we have a culture of innovation whereby it is not just "owned" by a single unit or entity, but extended to the entire organisation of more than 25,000 employees. More about that later, in Part 3 of this article.

First, let me start where my last article left off. After acquiring 8 AWS Certifications, I decided to pursue the AWS Specialty Certification in Machine Learning next. But I didn't want to just acquire this certification for the sake of it; I wanted to gain real knowledge and be able to solve real business problems with Machine Learning. So I decided to take a year off the "certification trail", to build up my knowledge before attempting to do the certification.

Dreaming Big

I've always feared delving into Machine Learning (ML). I wondered if I could take this certification on, without a PhD or a Math degree? I searched online, in vain, for learning resources simple enough for beginners that, while imparting knowledge, would allow me to gain hands-on experience in developing, architecting, and running ML / Deep Learning workloads in the Cloud.

Then I came across a workshop for AWS DeepRacer at the AWS Summit in Singapore, where we were taught to get hands-on with ML from the very start (regardless of skill level) - through a Cloud-based 3D racing simulator, fully autonomous 1/18th-scale physical race car driven by Reinforcement Learning, and a global racing league.

I was convinced that I had found the educational platform that I was looking for - one that took the fear and mystery out of learning ML, in a fun and gamified way!

It was a 2-day Summit, so after that initial workshop, I went back to train my own Reinforcement Learning model overnight, and took part in the DeepRacer race the very next day. To my surprise, I found myself in 2nd place on the leaderboard early that morning!

Encouraged by that initial success, I set myself a goal there and then.

BHAG: To learn about ML / Reinforcement Learning through AWS DeepRacer, compete in the global league, and become Top 3 in the World!

I ended up being 7th overall at the Singapore Summit, disappointed but determined to pursue my goal with conviction. So I spent the next 6 months honing my skills in Reinforcement Learning and DeepRacer, sleeping less than 4 hours on most nights (most of my learning was done at night after work).

I went on to compete at Amazon re:MARS (an AI conference held in Las Vegas) and 2 other Summits in Hong Kong and Toronto.

After two more heart-breaking 4th-placed finishes at Amazon re:MARS and the Hong Kong Summit, I finally made it onto the podium at the Toronto Summit! It felt amazing, because I clinched that 3rd-place finish under very dramatic circumstances, with less than 90 seconds left on the very last run of the day!

To top it off, that result meant that I scored enough points to secure the #1 spot on the global Summit Circuit leaderboard, surpassing my goal of being Top 3 in the world!

6 months into my Machine Learning journey, I had achieved my first milestone - one that many thought I was crazy to aim for. Indeed, it was a little crazy: to achieve this, I had travelled twice around the world, racing against brilliant minds in AI from across the globe! But in the end, it was all worth it.

Looking back, I now realise why I chased this trophy around the world so fervently - it was because I wanted to show to my son, that if you dream big about what you want to achieve, then go after it with hard work, determination and support from family and colleagues, you can be amongst the best in the world at what you do!

Cheers to all the Dreamers out there!

Contributing to the Community

A lot of what I learnt about Reinforcement Learning and AWS DeepRacer came from the DeepRacer Community. Being part of this community from its early days also opened my eyes to how a Community of Practice works best: when people with a common passion and goal interact, share and learn together, they exemplify the social nature of modern learning! I will share about getting involved in (and contributing back to) the community in Part 2 of this article.

Leading Innovation like an Entrepreneur

In the months following my mini-"victory" in Toronto, I graduallly came to the realisation that the innovation brought about by advances in AI/ML means that there needs to be a corresponding talent transformation in organisations which are (or will be) disrupted by such innovation. In Part 3 of this article, I explore how I became more entrepreneurial in my leadership style, as I began to share about my adventures with AWS DeepRacer to the rest of my organisation - along with the benefits that it brings to an AI/ML talent transformation.