SECTION 01
Principles
We optimize for systems people trust every day.
Every decision that is hard to reverse gets written down with the options we rejected.
- ClarityReadable code and plain names beat cleverness.
- OwnershipThe team that builds a service also operates it.
- EvidenceClaims about quality come with tests, metrics, or reviews.
SECTION 02
Delivery lifecycle
Work moves from issue to production in small, reviewable changes.
- PlanIssue with acceptance criteria and a definition of done.
- BuildShort-lived branch with tests alongside the change.
- ReviewAt least one reviewer; security-sensitive changes need two.
- ReleaseAutomated pipeline to preview, then production with rollback.
SECTION 03
Quality and testing
Tests run on every change. A red pipeline blocks the merge.
- Unit and integrationBusiness rules and data access covered by fast tests.
- End-to-endCritical user journeys tested in a real browser.
- AccessibilityAutomated checks plus manual keyboard and screen reader passes.
- PerformanceBudgets for page weight and response time, checked in CI.
SECTION 04
Security practice
Security work starts in the first sprint and never stops.
- AccessSSO with MFA, least privilege, quarterly review.
- SecretsStored in a managed vault, never in code or chat.
- DependenciesAutomated updates and vulnerability scanning.
- IncidentsDocumented response plan and post-incident reviews.
SECTION 05
Operations
Systems we launch are monitored from day one.
- MonitoringUptime, errors, and latency with alerts to an on-call engineer.
- BackupsAutomated, encrypted, and tested by restoring.
- Change logEvery production change is recorded with who and why.
SECTION 06
Default stack
We choose boring, well-supported tools unless there is a reason not to.
- FrontendReact, Next.js, TypeScript
- BackendNode.js, Python, PostgreSQL
- PlatformVercel, Supabase, AWS, GitHub Actions
- AIClaude and other models via audited integrations
