Smaller resources, larger leverage. AI Coding is the new lever.
Three early companies shaped how I work. A UK-regulated sportsbook taught me to ship across time zones under PMP-grade discipline, and to design under high-concurrency load. A multi-tenant SaaS taught me to scale by removing single points of trust. A medical compliance platform taught me to slow down when slowness is the right call.
In 2025 I started an experiment: stack those three lessons on top of an AI Coding workflow and see how far one person + the right setup can take commercial delivery. OPC is the shape of the experiment, not the goal.
I work well where there are real constraints, hard technical trade-offs, and complex business that needs breaking into shippable steps. Less well where the answer is more headcount, where there is no product opinion, or where "ship fast, fix never" is the culture.
These days I'd rather spend my time on the first list.
Reset from a 50-person org back to a one-person company in 2025, combining team-management instincts with the AI Coding workflow to redo what technical delivery looks like.
Made cross-level architecture calls in a medical compliance setting (e.g., skipping sharding and going straight to TiDB), landed cross-hospital imaging sharing in production — cumulative re-exam savings >¥30M for patients.
Lived through two failed business pivots, then sold the pain we'd just survived — as a fresh-supply SaaS for other operators.
Worked on middleware for high-concurrency bet placement at a lottery platform.
Led the SOA rewrite of SportsBook at a UK regulated sportsbook — first time fully experiencing international collaboration, PMP-grade discipline, and regulator certification.