AI Agent Architecture Articles
Technical articles on LLM agent architecture, orchestration patterns, memory boundaries, tool execution, state management, retrieval, and control-plane design.
Core articles
Parallel Exploration in LLM Systems Is an Orchestration Pattern
Why multi-path reasoning in LLM systems usually comes from inference-time orchestration rather than ordinary single-pass autoregressive decoding.
LLM-Led vs Orchestrator-Led Tool Execution
A control-plane placement comparison for tool-using LLM systems, covering reliability, observability, latency, cost governance, and security.
LLM Memory Boundary Model for AI Assistants
A vendor-agnostic model of context construction—what can enter context, what gets used per response, what is retained for later, and which security controls must live outside the prompt.
Human vs GenAI capability map (engineering view)
A practical mapping of human cognitive capabilities to GenAI limitations, engineering substitutes, and residual gaps.
Section resources
Context, reusable contracts, related links, and external baselines for this topic.
About this section About this section
Focus
- Orchestration patterns (control-flow mechanisms) — where control flow lives and how workflows are sequenced.
- State & lifecycle management — what persists across steps and boundaries (session/thread/run), what resets, and when.
- Tool invocation lifecycle — selection, authorization/enforcement, validation, error handling, retries, and egress constraints.
- Write paths — where and how the system can persist changes to external systems.
Terminology
- Orchestrator: the component that owns control flow (decides next action/tool call) and enforces policy.
- Model-led: the model proposes next actions/tool calls within constraints enforced by the orchestrator.