The actuary who quit and founded two artificial intelligence startups
Sam Zheng is a pioneer of artificial intelligence (AI) startups in Australia. Both of his startups, Hyper Anna and Curious Thing have been successfully funded by the likes of Airtree Ventures and Westpac’s VC arm Reinventure Group. But only 5 years ago Sam was an actuary at Suncorp and Zurich modelling earthquakes and other natural disasters, crunching numbers as an actuary and data scientist. We chatted to him about data science and AI, what it was like to start a startup, and starting an AI company in Australia.
On Data Science and AI:
Q: How do you define data science in layman’s terms?
It’s very broad, I think data science/AI value chain encapsulates a whole process. To deliver an AI solution we need algorithm engineers, data engineers, research scientists as well as designers and product managers. It should be seen as a package that delivers business value. Data science combines different disciplines and also addresses different types of problems. Using data to derive value existed for thousands of years but only has a trendy name now.
Lots of people say that data science is a recent concept. But we’ve always had it, engineers have used data ever since the disciplines of engineering were created. People have always used data to solve problems, especially in health and science. Without data, we wouldn’t have been able to come to conclusions or validated hypotheses — all of which shall start with the understanding of statistics, linear algebra and data interpretation.
All knowledge has come from some form of data.
I think the biggest difference in the modern information age is that data is created and analysed at a higher volume than ever before and at a crazy level of granularity — there’s data on everything!
Q: How does data science differ from data engineering?
I personally think that data engineering is a crucial subset of so-called data science. Data science is not just about modelling, it’s a combination of many different things. Only people without commercial experience would think that data engineering and data science are unrelated and mutually exclusive.
Q: Why do you think artificial intelligence has gained so much hype over the past few years?
The current hype started 5–6 years ago. But there has been hype about AI all through the 20th Century (e.g. in the 1970s and 1990s). In the past, the hype couldn’t be substantiated because the compatible computational and processing power wasn’t there to make AI a commercial technology. However, with better GPU (graphics processing unit), Deep Learning started to shine and delivered better outcomes in certain applications.
Specifically, the most recent hype has revolved around computer vision — which is still the most practical, ubiquitous, and advanced application for AI to date. Before Deep Learning, we were using tools like support vector machines for problems like image classifications. We could call this AI too. But the arrival of Deep Learning triggered an event which led to the current hype.
Q: No doubt AI can be applied to many business applications, why did you choose recruitment interviewing specifically?
We are in conversational AI vertical. Our underlying technology is not just built for recruitment. It’s the engine that asks purposeful questions to humans so we can readily collect and analyse the data from verbal responses. We use knowledge graph to map ontology to questions to deal with open-context environments — in our software the AI drives the questioning and conversation with the human. Recruitment is the first and most straightforward application of this technology.
Between the three co-founders at Curious Thing, we were inspired to follow the recruitment path and we have got endorsements from clients including AWS (Amazon Web Services). We think conversational AI and NLP (natural language processing), in general, will be the next computer vision — bringing true productivity and commercial outcomes to businesses.
We’re also in the works on something that isn’t recruiting focused. Watch this space!
Sam’s Career Advice:
Q: You started as an actuarial analyst at Suncorp and Zurich, eventually switching to data science at Quantium. What made you decide to change?
Quantium is also an actuarial company, I was still doing data and statistical modelling then. It was still data science/ML it’s just changed its name now and become sexier! Back then it was known as statistical modelling. I was a pricing actuary at Zurich for motor and fleet insurance.
What made me want to change? I like to do different things. I think many people will make career changes during their lifetime. I was driven by the desire to build an awesome product. Unless you do your own startup you don’t have that chance.
Ultimately, I wanted people in 50 years time to look back at what I had built and still be impressed with what I had done.
Q: Do you think data science will be one of the most popular jobs in the next decade or two? What advice do you have for students studying this?
To students studying it right now, I would advise them to have strong statistical knowledge, and focus on those aspects of their course.
I don’t know if data science will be the most popular degree. I would suggest all students, regardless of their degree, to be adaptive and ready for rapid change in the things around you (e.g. processes, technologies, environments).
The only thing that doesn’t change is change itself.
Data science is no doubt a good discipline, it builds great mindsets. But if you only focus on one programming language and assume that’s the core of your data science skills, that’s not enough. You need to build up a problem-solving mindset and be readily adaptable. Being able to program and think in a multi-faceted way as a researcher, engineer, and also a scientist is extremely important. But no doubt to be good at any discipline, technical skills are the bottom line.
I think it’s good that we are introducing new skills to these students. For example, GBM (gradient boosting machines) are something that students are now learning. The key thing is the world is changing and we should be ready.
Personal questions:
Q: In Australia and other Western countries, there are not many Asian males who become founders. What inspired you to pursue this path? Have you experienced any hardships due to this?
I actually think this is not true, it’s statistically Asian males and males from migrant backgrounds who are the most entrepreneurial. The founder of Zoom for example (Eric Yuan) is one example and he’s done well. To me, I wouldn’t even consider this as an attribute, whether I am male or female, Asian or non-Asian.
Did I face lots of difficulty or rejection? Yes. But I wouldn’t attribute my hardship to me being an Asian man. In my mind, people just didn’t want to buy my product. And I used this as a way to perfect my product, by always asking why and making sure the product is growing.
In this day and age, I think everyone is different in some way, you can label people if you like, but everyone is diverse in their own way. The key thing is that everyone is an individual. If we focus on the little details, you lose touch with the core problem you’re solving.
The biggest hardship I experienced had nothing to do with diversity, it was in facing my own limitations as a person and trying to constantly improve myself. It was difficult to realise (as an ambitious person) that you cannot change everything about yourself and you won’t be able to master all that you want to and that’s just life! You just have to move with it.
But this founder journey, realising who you are, is also the best experience. Would I be the best salesperson even if I put in 200% of my effort? No, I doubt it. But I know what I’m good at and I know how to build products and work with a great team. This realisation is tough but you just need to do it. I think that applies to everyone.
Keep discovering yourself!
Q: If you had all the money in the world what would you dedicate your life to?
I might still do what I do, I really enjoy this. I might spend more time on the product. I would have more freedom to build, because short term financial impact is no longer the most important thing, and I can build a product with less limitation and with more of a focus on big-picture concerns.
This piece was written by Jeffrey Yin of Textbook Ventures. We organise startup events, write newsletters and cater exciting activities for student entrepreneurs across NSW.
Sign up to our weekly newsletter and check out our Facebook to stay in the loop!
