Agentic AI Architect - Foundations and Design

Analyze agent requirements and select appropriate architectures (reactive, deliberative, BDI, ReAct, GOAP) based on task complexity, safety requirements, and explainability needs, justifying architectural decisions using established agent theory. 

Design and implement advanced RAG systems using techniques including Raptor, Splade, dense vs. sparse retrieval, and agentic RAG patterns, optimizing for enterprise-scale knowledge bases with proper chunking, embedding strategies, and retrieval orchestration. 

Specify agent goals, context, and reasoning patterns using symbolic and contextual goal representation, chain-of-thought frameworks, and structured decision trees that prevent agent drift and handle edge cases reliably. 

Implement comprehensive safety and alignment mechanisms including deterministic guardrails, sandboxing, circuit breakers, constitutional AI principles, access control (RBAC/ABAC), and reasoning trace capture for auditability. 

Design tool networks and multi-agent orchestration patterns with proper API contracts, error handling, and collaboration workflows (research → critique → refine, specialist ensembles, orchestrator hierarchies).