Multi-Agent LLM Systems
Foundations, patterns, protocols (MCP/A2A), memory, reliability, and cost of agentic systems.
9 posts · AI & LLMs
- 1
Agent Orchestration & Multi-Agent Systems — Suite Index
This document is a reading guide for an 8-part research series about how to use AI assistants (called "agents") to do complex work automatically.
- 2
Foundations: Agents, Agentic Loops & When to go Multi-Agent
What an "AI agent" actually is — and when it makes sense to build one. The word "agent" gets used loosely, so this doc gives you a clear way to think about it…
- 3
Context Engineering & Memory for Agents
How AI agents manage the information they hold in mind while working. Every AI has a limit on how much it can pay attention to at once — and jamming more in…
- 4
Applied: Agent Orchestration in Print-Flow-360
How the AI features inside Print-Flow-360 actually work right now — and where smarter AI could be added in the future.
- 5
Frameworks & SDKs Landscape
This document is a plain guide to the main "toolkits" (called frameworks) that developers use when building AI agents — software that can think through a task…
- 6
Cost, Performance & Economics of Multi-Agent Systems
How much it costs to run AI systems that use many "agents" (AI workers) at the same time, and how to keep that cost under control.
- 7
Reliability, Evaluation & Observability
Why AI agents that do long, multi-step jobs are hard to keep working reliably — and what you have to do to catch problems before they cause real damage.
- 8
Communication, Coordination & Interop Protocols (MCP, A2A, handoffs)
How AI "agents" (software helpers that do tasks on their own) talk to each other and to outside tools.
- 9
Orchestration Patterns & Topologies
This document is a catalogue of the different ways you can wire multiple AI "agents" together to get work done.