📚 DoorDash IT Knowledge Base
├── 🚨 Incident ManagementTask
├── RunbooksTask
├── P0 Incident Response Runbook [template]
├── Service Degradation Runbook [template]
└── Rollback Procedures [template]
├── Escalation PathsReference
└── On-Call Rotation & Contact Matrix
├── Jira Incident Ticket ConventionsReference
└── Post-Incident ReviewsConcept
├── 🔐 Security & AuthenticationTask
├── SSO / Okta Integration GuidesTaskNew
├── Okta SAML Integration — Sandbox Setup Guide
├── SAML Assertion Troubleshooting Reference
└── Credential Rotation Procedures
├── Authentication FlowsConcept
└── Security Incident RunbooksTask
├── 🏗️ Platform & ArchitectureConcept
├── Service CatalogReference
└── [Service Name] — Owner, Dependencies, Endpoints
├── System Design DocumentsConcept
├── API & SDK DocumentationReferenceNew
├── Internal API Reference — Endpoints, Auth, Rate Limits
└── SDK Integration Guides — Setup, Usage, Versioning
└── CI/CD Pipeline DocumentationTask
└── Jenkins Pipeline Runbooks & Deployment Procedures
├── 🚀 Engineering EnablementGolden Path
├── Golden Path — New Engineer OnboardingGolden PathNew
├── Day 1: Environment Setup
├── Week 1: First Deployment
└── Month 1: Service Ownership
├── Tooling Setup GuidesTask
└── Development StandardsReference
└── ⚙️ IT OperationsTask
├── Change ManagementTask
├── Deployment ProceduresTask
└── Monitoring & Alerting PlaybooksTask
🔎 ML-informed Retrieval — Search Parameter Specification
Standard keyword search was insufficient because engineers under incident pressure search by
intent ("how do I restart the auth service") not by document title ("SSO Service Restart Runbook v2.1"). I designed the full specification architecture for an ML-informed semantic search layer — Confluence's semantic ranking uses ML-based relevance scoring — and engineered the parameter logic to shape how that model routes queries to content. I authored the search parameter specification document as the engineering brief: defining each parameter's values, weighting rationale, and expected retrieval behavior. A platform engineer then implemented the parameters against the Confluence ML search layer, using my spec as the implementation reference. I validated retrieval accuracy against a test query library drawn from real incident reports, iterating on the parameter logic until retrieval met the 3-minute target.
🧭
Concept
Explains what something is or how it works. No action steps.
review: 180d · owner: arch team
📋
Runbook / Task
Step-by-step procedure for a specific operational task or incident response.
review: 90d · owner: responsible team
📊
Reference
Lookup tables, service catalogs, contact matrices, API specs.
review: 365d · owner: named DRI
🚀
Golden Path
Canonical onboarding path for a new engineer or service. Curated and endorsed.
review: 60d · owner: eng-enablement