
GitHub · GH-600AsociadoIngeniería de IA
Agentic AI Developer
Agent architecture, tool use, memory/state, evaluation, multi-agent coordination and accountability inside production SDLC workflows.
Los detalles del catálogo (alcance, notas, nombres de dominios) están en inglés, como los publican los emisores.
Datos del examen
Dominios del examen
- Prepare agent architecture and SDLC processes17.5%
- Implement tool use and environment interaction22.5%
- Manage memory, state, and execution12.5%
- Perform evaluation, error analysis, and tuning17.5%
- Orchestrate multi-agent coordination17.5%
- Implement guardrails and accountability12.5%
Weights are midpoints of the published ranges (15–20%, 20–25%, 10–15%, 15–20%, 15–20%, 10–15%) from the Microsoft Learn GH-600 study guide (page updated 2026-05-13), read 2026-11-03.
Para tener en cuenta
- Proctored 120-minute assessment; the issuer says interactive components may occur.
- IA durante el examen: Check the current exam resource policy. A coding-agent certification does not itself permit AI during the assessment.
Prepará con
Fuentes
Verificado el 27 oct 2026 contra el emisor. Los precios cambian y varían según el país. Confirmalos con el emisor antes de pagar.
Imagen de la insignia: la oficial del emisor
6 reportes de candidatos
Autorreportados. Sirven para ver cómo se prepararon, no para tasas de aprobación.
Microsoft's event page links this session as a technical roadmap for GitHub Agentic AI Developer preparation.
The author reports two unsuccessful attempts and says the actual assessment required more application reasoning than expected after the official modules. This is a counterexample to treating completion of the learning path as sufficient readiness.
A self-described successful candidate emphasizes building labs and reasoning about workflows, permissions, logs and implementation decisions. The useful preparation lesson is to practice actual engineering tasks rather than only reading definitions.
The author reports passing and highlights configuration of the software-delivery process and multi-agent orchestration. This supports investigating how agents operate within a development system rather than reducing the credential to prompting or code completion.
The author reports passing GH-600 among several assessments without dedicated advance preparation, while explicitly recommending that other learners prepare properly and complete hands-on exercises. This illustrates how short exam-specific review can exclude a substantial existing knowledge base.
The current listing specifies a 120-minute proctored assessment and role scope spanning agent architecture, tool interaction, memory/state, evaluation, coordination and guardrails. This is the primary reference for interpreting older candidate reports.
Certificaciones similares
- NVIDIA Agentic AINCP-AAI · US$200 · 2 h
- AWS Certified Generative AI Developer – ProfessionalAIP-C01 · US$300 · 3 h
- Anthropic Claude Architect ProfessionalCCAR-P · US$175 · 2 h
- Anthropic Claude Developer FoundationsCCDV-F · US$125 · 2 h
- LangChain Certified Agent EngineerLCAE · US$99 · 2 h
- NVIDIA AI Models and Application DevelopmentNCA-AMAD · US$125 · 1 h