Micro-Minds for Multi-Agent LLM Orchestration: Building Multi-Agent LLM Systems with Abhidhamma-Inspired Cognitive Decomposition
Format:
Paperback
En stock
0.31 kg
Sí
Nuevo
Amazon
USA
- Build multi-agent LLM systems you can actually trust.Most agent demos fail in production for the same reasons: ambiguous handoffs, drifting instructions, unverifiable claims, and “never-done” loops. Micro-Minds is a practical playbook for designing multi-agent LLM workflows as auditable systems—with explicit contracts, artifact-first state, and evaluation gates that stop bad work before it spreads.What you’ll learnHow to decompose complex objectives into stable agent roles and interfaces (inputs, outputs, preconditions, postconditions)How to externalize “memory” into versioned artifacts (schemas, rubrics, checklists, trace logs) so work stays reproducibleHow to route, validate, and terminate agent work using measurable gates (novelty, integrity, reliability)How to prevent drift, hallucinated evidence, and tool misuse from cascading across a workflowHow to design repair loops that improve quality without infinite recursionWhy it’s differentThe book borrows an Abhidhamma-style cognitive decomposition—not as religion, but as a disciplined way to model “what arises, under what conditions, and how it changes.” You’ll get a compact set of primitives you can reuse across extraction, synthesis, retrieval-augmented work, evaluation, and multi-agent coordination.Who this is forAI engineers, data engineers, and architects building agentic workflowsTeams shipping customer-facing LLM features that require audit trails and governanceBuilders tired of “prompt spaghetti” who want contracts, artifacts, and testsWhat you getReusable templates for agent briefs, artifact schemas, evaluation rubrics, and routing logicConcrete failure modes and countermeasuresA reference pipeline you can adapt to your domain
IMPORT EASILY
By purchasing this product you can deduct VAT with your RUT number