engineering-advancedlisted

# Engineering Advanced Advanced engineering patterns for AI-native startups building agents, RAG systems, APIs, and scalable infrastructure. ## Keywords Agent design, RAG, retrieval augmented generation, MCP, API design, REST, GraphQL, CI/CD, GitHub Actions, Docker, Kubernetes, microservices, event-driven, message queues, caching, database design, system design, observability, infrastructure as code, AI pipeline ## Core Domains ### 1. Agent Design **Agent architecture patterns:** | Pattern | Use Case | Complexity | |---------|----------|------------| | Single agent + tools | Simple tasks, clear workflow | Low | | Agent with sub-agents | Complex tasks, domain separation | Medium | | Agent team (orchestrator) | Multi-domain, parallel work | High | | Agent swarm | Autonomous exploration, research | Very High | **Agent design principles:** - Give agents clear, specific instructions (not vague goals) - Define tool boundaries (what the agent CAN and CANNOT do) - Implement guardrails (content filters, action limits, human-in-the-loop) - Design for failure (retry logic, fallback paths, error handling) - Observe everything (log prompts, responses, tool calls, latency) **Agent evaluation:** - Task completion rate - Average tokens per task - Tool call efficiency (fewer calls = better) - Error rate and recovery success - User satisfaction / output quality ### 2. RAG Architecture **RAG pipeline components:** ``` Documents → Chunking → Embedding → Vector Store → Retrieval → Gen