Founders agonise over the stack as if it were permanent. It is not — but the wrong early choices do create drag. Here is how we choose technology that still feels good in two years.
Optimise for boring where you can
Use proven, well-documented tools for the unglamorous 90%. Save your innovation budget for the part of the product that is genuinely novel.
Hire-ability matters
A niche stack with a tiny talent pool will haunt you. Popular, mature ecosystems mean easier hiring, more libraries and better AI tooling support.
Match the stack to the problem
SEO-critical site? Server-rendered framework, no question. Real-time collaboration? Pick infrastructure that makes it easy. Heavy AI features? Choose languages and platforms with first-class AI tooling.
Our defaults — Angular, Java/Spring, Node, Python, PostgreSQL on a major cloud — are not exciting, and that is the point. They are fast to build with, easy to hire for and they scale.
Marcus Lee
Principal Architect
Designs systems that scale. Obsessed with clean architecture, performance budgets and developer experience.