I build AI systems that make it into production, and I keep thirty-year-old computers running. The two habits feed each other more than you’d think.
Designing systems where AI carries real operational load, not just adding AI to a button. I use code generation, orchestration and agentic workflows to build capable products faster, automate complex decisions and extend what small teams can deliver.
Turning early ideas into scalable, well-structured products that survive contact with actual users. From requirements and architecture through to interfaces, integrations and deployment, I design the complete system around the problem it needs to solve.
Finding the parts of a business consuming hours of repetitive manual effort and replacing them with intelligent, reliable workflows. The goal is not simply to automate individual tasks, but to connect systems, remove bottlenecks and give people back meaningful time.
Bridging the gap between what a business wants to achieve and what technology teams can realistically build. I help organisations identify the right problem, choose an appropriate approach and turn technical complexity into a clear, practical path forward.
One foot in 486 builds, classic Macs and beige boxes that hum. The other in agents, on-chain weirdness and whatever AI shipped this week. Watching them rhyme is the fun part — every new idea was someone’s hack thirty years ago.
The machine that started it. Currently mid-recap on the bench.
Japan's arcade machine for the home. Still the best-looking case ever made.
Snow White industrial design, and it still keeps perfect time.
All-in-one 68040. The school computer, rescued and recapped.
Tiny console, enormous ambition. The interface unit is half the fun.
Arcade hardware in a living room box, with the price tag to match.
Fault-finding on a 1992 motherboard requires the same discipline as debugging a distributed system. Observe the behaviour, isolate the fault, test the theory and change one thing at a time.
Different scale. Same mindset. Just better light and a steadier hand.
Take the problem apart until only the physics of it are left.
The useful angle is usually the one nobody in the room has tried.
Build it small, put it in front of reality, change your mind quickly.
AI does the lifting. The judgement about what's worth building stays human.
Run your tools. Don't let them run you. The defaults were not designed for you.
The interesting answer is rarely the first one. Go east when the path goes north.
Deeply, then build boldly. Most things don't need more talking; some things need a lot more thinking.
Always open to a good conversation about technology — what’s working, what isn’t, and what’s worth building next.
mail@chrisfreeman.com.au