SKU/Artículo: AMZ-B0FX3DPF45

SEMANTIC KERNEL FOR AI AGENTS: BUILDING AUTONOMOUS SYSTEMS WITH LLMS : Design multi-agent workflows, tool use, and intelligent automation with Microsoft’s AI framework

Format:

Kindle

Kindle

Paperback

Detalles del producto
Disponibilidad:
Fuera de stock
Peso con empaque:
0.79 kg
Devolución:
No
Condición
Nuevo
Producto de:
Amazon
Viaja desde
USA

Sobre este producto
  • Build production-ready AI agents with Microsoft's Semantic Kernel: working code and battle-tested patterns without the guesswork.AI agents promise intelligent automation, but most tutorials leave you stranded between a demo and a deployable system. You need more than function calling basics. You need orchestration patterns that scale, retrieval strategies that stay grounded, and operational guardrails that prevent costly surprises in production.Semantic Kernel for AI Agents takes you from first principles to production deployment across .NET, Python, and Java. You'll design multi-agent workflows with clear routing logic, build tool catalogs that enforce least privilege, and wire OpenTelemetry so every decision is traceable. This isn't theory. Every pattern ships with runnable code, configuration templates, and operational playbooks refined in real environments.What You'll Master:Native function calling and tool design with typed schemas, validation, enums, and guardrails that prevent malformed inputsMulti-agent topologies using coordinator and worker patterns, with handoff strategies that keep capabilities separated and auditableRetrieval augmented generation with Kernel Memory pipelines, Redis and Elasticsearch hybrid search, and context injection that cites sourcesPrompt engineering for agents: system prompts as contracts, example calls that guide behavior, and versioning that survives model updatesOpenTelemetry integration for traces, metrics, and logs that answer "what did it do, how long, and how much"Offline evaluation with Prompt Flow, regression tests using transcripts and fixtures, and human review loops with clear acceptance criteriaContent Safety thresholds, secrets management per environment, OAuth with Microsoft Graph, and least privilege tool patternsAzure deployment strategies: Azure AI Foundry model lifecycle, Container Apps autoscaling, CI/CD with feature flags, and blue-green releasesPerformance tuning through token budgets, response caching, rate limit coordination, and concurrency controls that respect upstream quotasTroubleshooting playbooks mapping symptoms like slow turns, quota breaches, or grounding failures to stepwise fixesThree Complete Reference Builds:B2B support copilot with retrieval, routing, and safety gates that block unsafe content while citing policy documentsBatch enrichment pipeline processing thousands of records with idempotency, structured retries, and per-record cost trackingScheduling assistant using Microsoft Graph with OAuth, confirmation tokens, and guardrails that enforce business hours and external domain policiesEvery chapter includes working examples in .NET and Python, with Java coverage for core patterns. You'll see compact code for kernel registration, plugin schemas, adapter patterns for Azure OpenAI and GitHub Models, configuration that separates secrets from defaults, and telemetry that feeds dashboards your entire team can read. The book also covers advanced topics like provider neutrality, telemetry conventions mapping to OpenTelemetry fields, and drift detection that alerts when model behavior shifts unexpectedly.Whether you're migrating from template-driven planners to native function calling, wiring Kernel Memory with vector store connectors, or setting cost-per-task budgets that align with finance expectations, this guide delivers the practical architecture and operational discipline production systems demand.Stop reinventing the wheel. Grab your copy today and ship agents that scale, stay safe, and earn trust.

Fuera de stock

Selecciona otra opción o busca otro producto.

Este producto viaja de USA a tus manos en