Meet David Lewis — an Engineering Manager with 20+ years in tech and a long history in media companies. Lately, he’s been exploring AI-assisted development with tools like Claude Code — describing it less as “vibe coding” and more like Agile at light speed — while continuing to contribute to the Sydney dev community and meetups. We chatted with David for our Ducks in a Row newsletter.
Introduce yourself!
I’m David, Engineering Manager for Publishing Architecture at Nine Entertainment. I’m originally from the UK, but moved to Sydney 25 years ago this week. I was only supposed to come over for a year! I love running and chocolate, one enables the other.
What do you do for work?
I work within the Publishing area of Nine, that’s brands such as Sydney Morning Herald, The Age, AFR and nine.com.au. My role is brand new to the company, so I’m very excited to be helping forge what this role stands for. If it’s successful, our teams will be more confident taking on bigger challenges in an ever-changing technical environment. Right now, a big focus for my team is understanding how software development is changing in the face of the AI tooling now available.
What do you do outside work? Any hobbies? Projects?
I love the media; my original degree was very much centred around media and media production, but that was in the analogue age. Now I find that many of my hobbies centre around producing things, or how to produce things. Those things might be games, videos, apps. I get a lot of enjoyment from trying to figure out how to improve the crossover between technology and the production of the finished product.
This has led me to projects around sheep herding games, drawing apps, kick counters for when my partner was expecting our daughter, and an app to help people with their focus on reading. Not all of these are truly finished, but they’re always things I get in front of people, and they scratch a certain itch for me.
You’ve been sharing your thoughts on working with AI tools like Claude Code on Medium. You’ve described working with AI agents as “accelerated Agile” — what actually changes day-to-day for an engineer or team using this approach?
I feel that if I could answer that question well, I’d be a rich person! Day to day, what I think needs to change for a lot of software engineers is that they need to experiment and play with this technology to find ways of working that work for them. There is a tantalising prospect of being able to do more than you thought possible, but the processes to unlock this aren’t clear yet. It feels like the late 90s web where everything felt like such potential, but we hadn’t realised it yet - so we experimented and played, and eventually the processes emerged.
Right now, I think these tools have the potential to make us more ambitious with the things we take on, and I think in doing that, we should still be grounded in the things that made agile a good thing. Have a goal, create something that works, and get feedback on what you’ve made. I think those things aren’t new, but what I hope is that the loop gets faster and we can iterate on things more quickly to produce things that are highly valuable to people. Without that grounding, there’s a danger that these new ways of working will allow folks to run very quickly but without a focused direction.
As an engineering manager, how do you think AI is reshaping your role?
I’m a slow thinker, and that’s not a bad thing. I tend to want to spend time with things, dig around, and not have knee-jerk reactions to things. I think one of the things that AI is reshaping about myself is allowing me to be a little quicker than I used to be, to explore ideas a little quicker and to help me shape my thinking more clearly.
I’m also finding that it’s helping me get back to the technology with more confidence. I’m very much an EM that focuses a lot on the people aspect of the role, to create teams that people look back on and draw inspiration for their future careers. That’s my foundation, and in the past, I’ve struggled to balance that type of EM with continuing to stay relevant on the technology as it always evolves. AI has helped me return to my tech roots and rebuild my perspective, to then go on to further support the teams I manage - I quite like that.
What’s something AI still gets wrong (or struggles with) that engineers shouldn’t blindly trust?
I believe that engineers still need strong foundational skills, and the reason for that is that sometimes the AI can enter a bit of a spiral. When the solution isn’t quite clear, I find that it starts reaching out into places it shouldn’t, trying to make changes in places that may not be necessary. When that starts happening, I’ve seen it burn through 1000s of tokens, and I have to stop it and suggest a direction myself, then the solution comes a little clearer and more quickly.
So I would say don’t blindly trust it if it appears to be doing this. When it’s taking a long time to do something and making a lot of changes, ask it to explain its thinking and reset.
You built a meal planning app from a real-life problem — do you think AI is pushing us back toward more problem-first, less spec-heavy development?
The meal planning app was me exploring some techniques for developing an app. With that, I was exploring two ideas: firstly, that user stories offer a great way to communicate with an AI, and secondly, that you can rely on automation to ensure things stay under control. What I’m hoping is that there’s a world where we have small teams with folks focused on problems, quickly iterating on those ideas and getting quality feedback. I hope we don’t regress into a world where people try to spec out things upfront and expect to one-shot everything - I feel the interesting things happen when people use things and ideas form from a rich set of perspectives.
Any podcast you are currently listening to, or a book you are reading?
I listen to Pivot and The Rest is Politics most weeks, currently re-reading a couple of books and trying to see how their ideas apply in 2026 (Team Topologies and An Elegant Puzzle). Just finished Enshittification - as I wanted to know more about the term.
What is your favourite tool or resource, and why?
It’s hard to go past Claude Code at the moment; it’s pretty mind-blowing what that tool can do, and I enjoy it most when I feel that it’s amplifying my own skills and ideas. Currently a bit obsessed with how web front-end fundamentals can be a strong foundation for creating UI without cutting corners.
Your one-sentence work-related advice
Stay curious and have fun with technology.
You might also enjoy these interviews:
Kristine Howard: Lessons in Leadership, Making Your Own Luck, and Retiring Early
Inga Pflaumer, Head of Engineering at Relevance AI
Subscribe to our monthly newsletter to get our latest content in your inbox.
