Next to take his place in the hot seat is Consultant Technologist, CTO and Executive Advisor. Welcome, Vlad! We hope you’re as excited as we are to be a part of this series. So without further ado...
Q: What excites you most about the future of your industry?
A: How prevalent technology continues to become and how low the barrier for entry keeps getting. In the 90’s, when I made first contact with computers, working with them was in many companies a domain separate from everything else. Even CAD or accounting software experts carried the “computer person” label. Nowadays computers and software are basic tools that seamlessly integrate into our workflows. A good example is how proficient people who do not necessarily have a data background are in data analysis software and use it in operations, marketing, sales, etc. We continue to discover new problem spaces where technology fits and build solutions for them, which in turn creates new markets.
Q: Which technological innovation will drastically change businesses in the next five-to-10 years?
A: Machine learning has tremendous potential. But to achieve that it needs to be truly democratised. Cutting edge research is being done and productised in a select few places, but the majority of companies are followers and consumers of that. There are very real cost and vendor lock-in concerns around using ML on large data sets in the cloud, as highlighted by Andreessen Horrowitz in a February 2020 report. I think the future for mass machine learning is across a multitude of diverse environments, especially at the edge, where data is generated. Technology that is (cost) optimised for specific domains enables that. We already see early adoption in phones, with automotive, healthcare, energy, public transport, shipping and logistics, image & sound processing and civil engineering following suit.
Another challenge is acquiring and analysing the correct data. A recent study published in the British Medical Journal (BMJ 2020;368:m689) examined 91 medical imaging deep learning clinical trials. A majority of them had a high risk of bias and gaps in methodology (randomised vs non-randomised), data gathering (prospective vs retrospective), adherence to reporting standards, real world clinical testing and code availability and transparency. Yet 61 studies claimed their ML performance was at least comparable to or better than that of clinicians, while only 31 studies stated that further prospective studies or trials were required. More people from every discipline (sciences, technology, legal/compliance) meeting in the middle, working together to close these gaps should address that.
Q: What’s the best piece of advice you’ve been given?
A: Best advice - think of how others perceive your words and actions. I received this some years back when as an enthusiastic technology leader, I was building a high performing team and a complex product at the same time. Operating at breakneck speeds often means cutting some corners, and when that impacts your communication style (too often, too seldom, too verbose, too succinct for example) it is easy for people to misinterpret your thoughts and intentions.
Worst advice - outsource engineering to reduce cost (I didn’t follow it).
Q: If you could tell your younger self one thing, what would it be?
A: Slow down, catch your breath, you cannot do it all in one day. Pretty cliché, right? But it is true.
Q: Name a life-changing book/podcast you’d recommend to others.
A: Without a doubt, Ben R. Rich’s: ‘Skunk Works: A Personal Memoir of My Years at Lockheed’. I absolutely despise war and warfare, but such was the Cold War era Rich lived in (another prime example of how technology evolves ahead of society). It is one of the best books on engineering, leadership and management I have read, written by the man who contributed to the U2 and designed and built the SR-71 and F-117 planes. Besides talent and hard work, he achieved that with cross-functional teams, short feedback loops and other organisational niceties we as engineers appreciate, long before the terms we now use to describe them were even coined.
Q: What would you like to be remembered for?
A: Having built genuinely useful products while helping others become better at what they do.