
From the team
Notes on AI, engineering practice, and what it takes to ship production software.
Shipping AI features that survive production
A demo that works once isn't a feature. Here's what changes when a model has to hold up under real traffic, real data, and real cost pressure.
View detailsThe multi-tenant mistakes that show up in year two
Most SaaS architecture problems don't show up at launch. They show up once you have real customers with real data volumes and real support tickets.
View detailsStaff augmentation without losing institutional knowledge
Embedded engineers can move fast without leaving your team worse off when the engagement ends. It depends on how the engagement is structured from day one.
View detailsWhat ISO/IEC 27001 actually changes day to day
The certificate is the least interesting part. The habits it forces, and keeps forcing after the audit is over, are what actually reduce risk.
View detailsWhen Web3 is the right answer, and when it's not
A blockchain is the right tool for a narrow set of problems. Most products that ask for one actually need something simpler.
View detailsHow we scope a free 15–30 minute working session
Fifteen to thirty minutes is not enough to review a whole codebase. It is enough to learn whether your biggest risk is technical, organizational, or somewhere in between.
View details