The rise of AI has transformed not just what we build, but who gets to build it. With coding becoming more accessible than ever, traditional pathways into tech are being reshaped by creators who combine intuition, experimentation, and curiosity to unlock powerful innovation.

Luminance’s AI Product Engineer, Joe Danbury, is one of those trailblazers: a ‘vibe coder’ who followed a non-conventional route into the AI world. We sat down with him to discuss how he joined Luminance, what it's like to build cutting-edge features from scratch, and his advice for anyone wanting to break into the field.

What were you doing before you joined Luminance?

I studied Philosophy at the University of Bristol. It might not sound like a typical AI background, but it taught me critical thinking and logic – skills I use every day in my role now.

I didn’t have a set career plan but after applying for some grad jobs I started working at a small financial PR firm. That’s where I first got to experiment with AI in a real-world setting, which planted the seed for everything I’m doing now.

How did you first get into the AI space?

Like many others, when ChatGPT first came out I was super excited by the possibilities it brought and was playing around with it for a lot of random things, from creative writing to life admin. Then when I entered the world of work, I saw just how many processes in business could be improved with AI.

I remember when we had to disable our AI notetaker during a sensitive call – I realized there must be a way to get around this. So, I built a solution. It auto-detected sensitive info, encrypted it, sent redacted text to an LLM, and preserved context well enough for coherent responses. I had no coding experience apart from a failed go at Python, but I kept pushing through. It certainly wasn’t plain sailing but the first time it properly worked it was magical and I knew I wanted to do this more.

What made you join Luminance?

I happened to meet someone who worked here at an event, and they described the company in a way that immediately clicked with me. There’s so much noise in AI right now and many companies are just riding the hype wave. It was clear to me that Luminance isn’t one of them.

They’ve been working on cutting-edge, proprietary AI long before it was trendy. They genuinely innovate, and that aligned perfectly with how I like to work: experimenting, building, iterating. I knew this style of working would suit me and I’ve been proven right since day one.

What’s your role like now?

I’m in the Product team but work closely with a whole range of teams, from users to commercial leaders to developers. The role itself is very self-directed; I get given abstract problems and then am just set loose to research possible solutions and ultimately design and build something that actually works. So, there’s a lot of freedom and plenty of chances to be creative which I really enjoy.

Often ‘vibe coders’ are thought of as ignoring or giving up upon finding bugs. I certainly do spend a lot time solving bugs but I see that as a key part of the job – and it’s very satisfying working out what’s going wrong and then being able to fix it yourself.

What are the most important skills for your role?

On the technical side, I'm working with LLMs day in, day out. Understanding how they work, their limitations, how different models perform against others, the scope of tasks I can perform using them, and how to safeguard my work against their faults, is key. Knowing how code works is useful - whilst I don't need to know how to manually write an agent from scratch, debugging gets much easier if I know how the concrete, deterministic code corresponds to the abstract plan I'm trying to implement.

Then there are the soft skills: curiosity, creativity, persistence. Things won’t work first time and things you thought would work might break. You need the grit to keep going, and the people skills to take feedback and improve your work, whether that’s presenting to stakeholders or taking on feedback from users. And ultimately, logic is everything. Whether it’s planning a project or hunting down a stubborn bug, logical reasoning is the foundation.

Any advice for people who want to do what you’re doing?

1. Build as much as you can. Make projects that solve problems you care about in your day to day life, whether it’s a personal pain point or a business issue. It doesn’t have to be perfect. The goal is not to build the next unicorn (but if you can do that, great!), but to showcase that you have the drive and self-motivation to put your energy into the sort of thing you'd want to be doing at work.

2. Own your time. You don’t need to work 24/7, but you need to be disciplined. Even one hour a day is enough to make huge amounts of progress. But if you start making excuses, you start falling behind.

If you're curious, enjoy solving problems, and are ready to be challenged, Luminance is the place. Inspired? We're hiring!