Engineering Standards

Updated Oct 20, 2026

Engineering Standards

How we build and ship software at NinjaTrader — the baseline every engineering team builds on, so new work starts from what's already settled.

Read the standard

Owned by the Engineering Navigators·Reviewed quarterly·exceptions via ADR·changelog

What it covers

Language lanesScala, Python, TypeScript, Dart, C# — the lane follows how failure looks.Four-eyes reviewIndependent approval on every production change. No exception path.Quality gatesNothing merges red. Gates run in CI, never on someone's machine.Auth & securityThe platform is the identity provider. No long-lived credentials.

Browse the standard

31 sections

Team standardsLanguage, framework, deploy and tooling per surface.PrinciplesWhich lane a new service belongs in, and why.ArchitectureLayering, generated contracts, incremental migration.Four-eyesIndependent review on every production change.Code ownershipEvery repository has an owning team.Audit responsibilitiesEvidence as a by-product of the work.Quality gatesNothing merges red.TestingUnit, integration, e2e, security, load, accessibility.CI/CDBuild once and promote.AuthenticationThe platform is the identity provider.ObservabilityErrors, telemetry, logs, dashboards, paging.SecurityNo long-lived credentials; least privilege.Feature flagsShip dark, roll out gradually.Design docs, RFCs & ADRsWrite one when it earns its place.Release & incidentsProgressive exposure; retro every incident.InternationalisationEvery surface ships translatable.DocumentationREADME, AGENTS.md, runbooks.LicensingPermitted licences and who to ask.When it doesn't fitConsult, then record the exception as an ADR.AI readinessThe map for making a repository ready for coding agents.Agent instructionsDurable repository rules for coding agents.Repository contextArchitecture, ownership and boundaries agents need.Development environmentA reproducible path from clone to running code.Task automationDiscoverable, non-interactive commands for routine work.Tools and integrationsApproved assistants, connectors and local tools.Testing and feedbackFast checks with clear, deterministic output.Security and dataLeast privilege, secret handling and data boundaries.AI change workflowTask intake, review and accountability for AI-assisted changes.AI maintenanceOwnership and upkeep for agent-facing repository content.Open questionsWhat is not yet settled.ReferencesControlling policy and templates.