Johnson Partners Diversity Series – Seiji Armstrong
Winner of the Community & Advocacy/Not-for-Profit category in the 2021, 40 Under 40 Most Influential Asian-Australian Awards is Chinese-Australian, Frances Voon.
Now a cyber security specialist leading a dedicated team at Google’s headquarters to protect child safety and eradicate Covid-19-related racial vilification online, Seiji Armstrong is on a mission to make the internet safer – a task that’s never been more daunting than during the pandemic. Johnson Partners (JP) caught up with him to discover what triggered his shift in focus from quantum physics to building algorithms to stop cyber abuse, discovering it’s an inexhaustible quest.
JP You obtained a PhD in quantum physics and during the course of your studies at the Australian National University (ANU) were a recipient of the Prime Minister’s Australia-Asia Award for outstanding research. Using your award funds, you transferred to the University of Tokyo where you led a team to develop a cluster of more than 10,000 quantum systems (the world record at the time was 14). How has your research been beneficial to the cyber world?
The experiment was a proof of concept, validating ideas that quantum theorists had shown were possible. What we built wasn’t useful as a platform, but there are at least two serious start-ups in Silicon Valley today who are using some of the same ideas (temporal multiplexing) to build a useful quantum computer.
JP After setting up global strategy at the Centre for Quantum Computation and Communication Technology for a year following completion of your PhD, you moved to Silicon Valley where you joined a cyber security start-up and spent three years building a world-class data team who developed AI-powered cyber security solutions that protected several Fortune 100 banks and retailers. What prompted this new direction, combining the intersection of big tech and global governance?
I had always been fascinated by how the internet works, and there is no better way to learn how the various pieces work together (or are sticky-taped together) than building cyber security solutions for large players who rely on the web. The move was motivated by the much larger platform offered by Silicon Valley, from which to affect positive change. The intersection of big tech and governance still fascinates me, and I’m much more optimistic than I used to be.
“There will unfortunately always be security — as long as there is motivation (financial or otherwise) for evading policies. Today a lot of the work revolves around influencing various teams with different incentive structures to work together to build robust systems that follow present day ‘engineering best’ principles. ”Seiji Armstrong
JP Your big career jump came in 2019 when you moved to Google where you remain as the Head of Machine Learning, Trust and Safety Organisation. You and your team have developed protections across a wide range of verticals such as violent extremism, child safety (and more recently Covid-19 related racism). Joining Google a year before the pandemic arrived, how different was your focus and operation then compared to now?
I had a fairy tale year in 2019 at Google. The campuses are well set up — I was playing in two lunchtime basketball teams plus a soccer team, and spent an evening a week breakdancing. This was a great way to connect with colleagues outside of work, and it makes work fun. A lot of those friendships and the trust we’d built up together made the video meetings during Covid years go fairly smoothly. The challenge has been building rapport with new team members whom I haven’t met in real life — especially as people are now distributed.
JP Essentially, you develop machine-learning algorithms to pick up on online abuse or content that violates policies. In 25 years, how different will this career path be and what do you foresee to be the largest looming cyber threats?
There will unfortunately always be security — as long as there is motivation (financial or otherwise) for evading policies. Today a lot of the work revolves around influencing various teams with different incentive structures to work together to build robust systems that follow present day ‘engineering best’ principles. The algorithms, policies, and code itself is secondary. I imagine how we work with one another will look drastically different in 5-years, let alone 25-years. Any answer I give on looming threats will turn out poorly, or worse, give people ideas. No comment!
JP Stepping back in time, you were born in Australia but moved to Japan at the age of three, learning Japanese as your first language. You returned to Australia at 8, not speaking English at school to begin with. Living abroad again, how do you define your nationality now compared to when you were a child, and how important is it that you feel a sense of belonging to a place?
Being bullied in Japanese pre-school for being white, and bullied in Australian primary school for being someone who must be asked “where are you from?”, you learn to downplay your differences. I would convince myself that belonging to another culture was a burden — and kids would reinforce this idea. Then you grow up and learn to harness it as a super power. Having more than one culture to draw on allows you to critically examine each. I don’t define myself by my nationalities — I learn to navigate societies that still insist on doing that for me.