Skip to main content
COMPEL Body of Knowledge
Get Certified Platform

COMPEL Specialization Stream · AITE-SAT

COMPEL Academy — AI Solutions Architect Expert

Expert certification credential for solution architects who design enterprise-grade AI solution architectures aligned with the COMPEL lifecycle, responsible-AI controls, and MLOps best practices.

Expert Technical Track 60 hours 60 CE credits

Profession title: AI Solutions Architect Expert

Audience: Solution architects, platform engineers, and technical leads designing enterprise AI solutions.

AI Solutions Architect Expert badge

Enroll in the AITE-SAT track

Registration, enablement, and the proctored assessment are delivered through compel.one. Seats open continuously.

Register on compel.one →

Prerequisite chain

  1. AITF AI Transformation Foundations
  2. → AITE-SAT (this credential)

Learning outcomes

The learning journey is sequenced to cover each outcome below in order. Every article in the journey maps to at least one outcome.

  1. 1. Design a COMPEL-aligned reference architecture for an AI use case.
  2. 2. Select LLM, RAG, and agent patterns under governance constraints.
  3. 3. Define non-functional requirements that satisfy audit and operations.

Body of Knowledge articles (48)

Module M1.1 (48 items)

  1. Article
    The Enterprise AI Reference Architecture
    M1.1-Art01
  2. Article
    Model Selection Decision Framework
    M1.1-Art02
  3. Article
    Prompt Architecture: Templates, Versioning, Injection Defense
    M1.1-Art03
  4. Article
    Retrieval-Augmented Generation: When, Why, How Much
    M1.1-Art04
  5. Article
    Chunking and Embedding Strategy
    M1.1-Art05
  6. Article
    Vector Stores: Selection, Hybrid Retrieval, and Reranking
    M1.1-Art06
  7. Article
    Tool Use, Function Calling, and Agent Loops
    M1.1-Art07
  8. Article
    Model Serving Patterns and Inference Paths
    M1.1-Art08
  9. Article
    Inference Cost Architecture: Caching, Routing, and Distillation
    M1.1-Art09
  10. Article
    Fine-Tuning Decision Tree: RAG → Few-Shot → PEFT → Full Fine-Tune
    M1.1-Art10
  11. Article
    Evaluation Architecture: Offline, Online, and Human
    M1.1-Art11
  12. Article
    LLM-as-Judge and Human Review Pipelines
    M1.1-Art12
  13. Article
    Observability for AI Applications
    M1.1-Art13
  14. Article
    Security Architecture for AI Applications
    M1.1-Art14
  15. Article
    Data Pipeline Architecture for AI
    M1.1-Art15
  16. Article
    Multi-Tenancy in AI Systems
    M1.1-Art16
  17. Article
    Latency, Cost, and Scalability Architecture
    M1.1-Art17
  18. Article
    Deployment Topology and Data Residency
    M1.1-Art18
  19. Article
    Environment Promotion and Change Management
    M1.1-Art19
  20. Article
    SLO, SLI, and Incident Response for AI
    M1.1-Art20
  21. Article
    Model, Prompt, and Index Registries
    M1.1-Art21
  22. Article
    Regulatory Mapping — EU AI Act Articles 9-15 for Architects
    M1.1-Art22
  23. Article
    Architecture Decision Records and Documentation
    M1.1-Art23
  24. Article
    Architecture Runway: Building the AI Platform
    M1.1-Art24
  25. Article
    Legacy Integration: Calling AI from CRM, ERP, EHR, Mainframe
    M1.1-Art25
  26. Article
    Build vs. Buy vs. Integrate
    M1.1-Art26
  27. Article
    Multimodal Architecture: Vision, Audio, Document
    M1.1-Art27
  28. Article
    Architecture Review Gate: Calibrate and Organize Stages
    M1.1-Art28
  29. Article
    Architecture Review Gate: Model and Produce Stages
    M1.1-Art29
  30. Article
    Architecture Review Gate: Evaluate and Learn Stages
    M1.1-Art30
  31. Article
    Responsible-AI Architecture Patterns
    M1.1-Art31
  32. Article
    Architecture for Agentic Use Cases
    M1.1-Art32
  33. Article
    Cost Model and FinOps for AI
    M1.1-Art33
  34. Article
    Architecture Handoff and Operating Model
    M1.1-Art34
  35. Article
    Capstone: A Complete Reference Architecture Package
    M1.1-Art35
  36. Lab
    Lab 01: Design a RAG Reference Architecture for a Regulated Internal Knowledge Assistant
    M1.1-Art51
  37. Lab
    Lab 02: Build an LLM Evaluation Harness with Offline, Online, and Human Components
    M1.1-Art52
  38. Lab
    Lab 03: Architect an Agentic Trading-Desk Assistant with Safety and Observability
    M1.1-Art53
  39. Lab
    Lab 04: Design a Secure LLM Gateway with a Policy Engine
    M1.1-Art54
  40. Lab
    Lab 05: Red-Team a Production LLM Feature Using the OWASP LLM Top 10
    M1.1-Art55
  41. Case Study
    Case Study: Morgan Stanley Wealth Management and the Internal-Assistant Rollout
    M1.1-Art61
  42. Case Study
    Case Study: BloombergGPT and the Domain-Specific Fine-Tune Decision
    M1.1-Art62
  43. Case Study
    Case Study: Harvey AI and the Legal Enterprise Deployment
    M1.1-Art63
  44. Template
    Artifact Template: AI Solution Architecture Design Document
    M1.1-Art71
  45. Template
    Artifact Template: LLM Evaluation Harness Specification
    M1.1-Art72
  46. Template
    Artifact Template: RAG Data Contract
    M1.1-Art73
  47. Template
    Artifact Template: LLM Gateway Policy
    M1.1-Art74
  48. Template
    Artifact Template: Agentic Runtime SLO and SLI Sheet
    M1.1-Art75

Competencies demonstrated

  • → Reference architectures for enterprise AI
  • → LLM + RAG + agent patterns with governance
  • → MLOps and platform engineering for AI
  • → Non-functional requirements for AI systems

Exam blueprint summary

Assessment
Proctored examination
Passing score
75% passing score
Portfolio
Required
Renewal
Every 24 months
Recommended hours
60
CE credits
60

Linked Core Mastery context

The Specialization Stream assumes AITF Foundations fluency. These Core Mastery resources are the recommended grounding before entering the AITE-SAT learning journey.

Formal credential definition

The machine-readable Open Badges 3.0 / W3C Verifiable Credential definition for AITE-SAT is published at /credential/aite-solution-architecture . HR platforms and AI citation engines can fetch the JSON-LD document at /credential/aite-solution-architecture.json .