Certified Artificial Intelligence Harness Engineer

Table of Contents

868 pages · 1 Day, 4 Hours, 22 Minutes of Read Aloud audio

  1. FRONT COVER MATTER
  2. WHAT ARTIFICIAL INTELLIGENCE (AI) IS AND IS NOT
  3. HOW LLMs WORK FOR THE NON-SPECIALIST
  4. FROM MODEL TO APPLICATION
  5. THE THREE DISCIPLINES AROUND THE MODEL
  6. WHAT A HARNESS IS
  7. WHY HARNESS ENGINEERING EMERGED AS ITS OWN FIELD
  8. THE SEVEN-LAYER ETCLOVG TAXONOMY
  9. THE ANATOMY OF A PRODUCTION HARNESS (COMPONENT MAP)
  10. THE CORE AGENT LOOP
  11. STATE AND CONTROL FLOW
  12. CONTROL STRUCTURES AND DECISION ROUTING
  13. RUNTIME EXECUTION ENVIRONMENTS
  14. WHAT TOOLS ARE
  15. THE TOOL REGISTRY
  16. DESIGNING GOOD TOOLS
  17. THE MODEL CONTEXT PROTOCOL (MCP) AND TOOL STANDARDS
  18. COMMON TOOL CATEGORIES
  19. CONTEXT AS THE SCARCE RESOURCE
  20. CONTEXT ROT AND CONTEXT BLOAT
  21. CONTEXT CONSTRUCTION (THE CONTEXT BUILDER)
  22. COMPACTION AND SUMMARIZATION
  23. MEMORY HIERARCHIES FOR CONTEXT
  24. STARTING A RUN
  25. LIFECYCLE HOOKS
  26. BUDGETS AND RESOURCE LIMITS
  27. ENDING A RUN
  28. WHY OBSERVABILITY MATTERS
  29. TRACING AND SPANS
  30. LOGGING, METRICS, AND DASHBOARDS
  31. REPLAY AND DEBUGGING
  32. THE EVALUATION HARNESS
  33. DESIGNING EVALUATIONS (EVALS)
  34. SCORING METHODS
  35. BENCHMARKS AND LEADERBOARDS
  36. TESTING THE HARNESS ITSELF
  37. THE SAFETY ARCHITECTURE
  38. PERMISSIONS AND ACCESS CONTROL
  39. PROMPT INJECTION AND TOOL POISONING
  40. MODEL CONTEXT PROTOCOL (MCP) AND SUPPLY-CHAIN SECURITY
  41. SANDBOXING AND ISOLATION FOR SAFETY
  42. RESPONSIBLE AI, COMPLIANCE, AND RISK
  43. WHY USE MORE THAN ONE AGENT
  44. ORCHESTRATION PATTERNS
  45. COORDINATION MECHANICS
  46. COMMUNICATION AND PROTOCOLS BETWEEN AGENTS
  47. TYPES OF AGENT MEMORY
  48. RETRIEVAL-AUGMENTED GENERATION (RAG)
  49. KNOWLEDGE MANAGEMENT FOR AGENTS
  50. SURVEY OF AGENT FRAMEWORKS
  51. STANDARDS AND CONVENTIONS
  52. PLATFORMS AND MANAGED SERVICES
  53. THE OPEN-SOURCE LANDSCAPE
  54. THE ECONOMICS OF AGENTS
  55. COST OPTIMIZATION TECHNIQUES
  56. LATENCY OPTIMIZATION
  57. RELIABILITY AND ERROR RECOVERY
  58. FROM PROTOTYPE TO PRODUCTION
  59. DEPLOYMENT ARCHITECTURES
  60. AGENT OPERATIONS (AGENTOPS)
  61. INCIDENT RESPONSE FOR AGENTS
  62. CODING AGENT HARNESSES
  63. COMPUTER-USE AND BROWSER AGENTS
  64. RESEARCH, WRITING, AND KNOWLEDGE AGENTS
  65. DOMAIN-SPECIFIC HARNESSES
  66. PROJECT ONE - A MINIMAL HARNESS
  67. PROJECT TWO - ADDING CONTEXT MANAGEMENT
  68. PROJECT THREE - TOOLS, THE MODEL CONTEXT PROTOCOL (MCP), AND SAFETY
  69. PROJECT FOUR - EVALUATION AND OBSERVABILITY
  70. PROJECT FIVE - MULTI-AGENT AND PRODUCTION HARDENING
  71. THE ROLE OF THE AI HARNESS ENGINEER
  72. PROFESSIONAL PRACTICE AND COLLABORATION
  73. ETHICS AND SOCIETAL IMPACT
  74. PREPARING FOR CERTIFICATION
  75. APPENDIX A: GLOSSARY OF TERMS AND ACRONYMS
  76. APPENDIX B: THE COMPONENT REFERENCE
  77. APPENDIX C: PATTERNS CATALOG
  78. APPENDIX D: SAFETY AND SECURITY CHECKLISTS
  79. APPENDIX E: EVALUATION TEMPLATES
  80. APPENDIX F: FURTHER READING AND RESOURCES
  81. APPENDIX G: LAB SETUP GUIDE
  82. APPENDIX H: RESUME PORTFOLIO BUILDING CAPSTONE PROJECTS
  83. BACK COVER MATTER