ai-engineering-from-scratch/phases/14-agent-engineering/15-crewai-role-based-crews/quiz.json at main · Amjad-AbuRmileh/ai-engineering-from-scratch · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
{
  "lesson": "15-crewai-role-based-crews",
  "title": "CrewAI: Role-Based Crews and Flows",
  "questions": [
    {
      "stage": "pre",
      "question": "What are CrewAI's four primitives?",
      "options": [
        "Agent, Task, Crew, Process",
        "Node, Edge, Reducer, Checkpointer",
        "Actor, Message, Inbox, Runtime",
        "Tool, Prompt, Model, Memory"
      ],
      "correct": 0,
      "explanation": "Agent, Task, Crew, and Process are the four primitives."
    },
    {
      "stage": "pre",
      "question": "What is the recommended production starting point per the CrewAI docs?",
      "options": [
        "Crew",
        "Flow",
        "Sequential process",
        "Hierarchical process"
      ],
      "correct": 1,
      "explanation": "The docs say start production apps with Flows; fold Crews in as sub-steps when autonomy earns its cost."
    },
    {
      "stage": "check",
      "question": "What is the difference between a Crew and a Flow?",
      "options": [
        "Crew is paid, Flow is free",
        "Crew is autonomous and LLM-driven; Flow is event-driven, code-owned, deterministic and testable",
        "Crew runs on GPUs only",
        "Flow has no agents"
      ],
      "correct": 1,
      "explanation": "Crew is autonomy-first; Flow is determinism-first."
    },
    {
      "stage": "check",
      "question": "Which of these is NOT one of the four memory types CrewAI ships?",
      "options": [
        "Short-term",
        "Long-term",
        "Entity",
        "Quantized"
      ],
      "correct": 3,
      "explanation": "Short-term, long-term, entity, and contextual are the four. Quantized memory is not a CrewAI concept."
    },
    {
      "stage": "check",
      "question": "What is backstory bloat and how does the lesson recommend handling it?",
      "options": [
        "Storing too many crews; archive old ones",
        "2000-word agent backstories push out context budget; keep them tight",
        "Too many tools; delete some",
        "Slow Bedrock calls; switch regions"
      ],
      "correct": 1,
      "explanation": "Backstories shape tone and judgment but eat context if oversized."
    },
    {
      "stage": "post",
      "question": "When is Hierarchical process worth picking over Sequential?",
      "options": [
        "Always",
        "When you have 4+ specialists that need a manager Agent to route between them",
        "When you have one task",
        "When you do not have any tools"
      ],
      "correct": 1,
      "explanation": "Hierarchical adds a manager Agent; only worthwhile when several specialists need dynamic routing."
    },
    {
      "stage": "post",
      "question": "Why does the lesson caution against using a free-form Crew in production?",
      "options": [
        "Crews are unsupported",
        "Output variability is high and debugging is painful without a Flow wrapper",
        "Crews cost more than Flows",
        "Crews lack memory"
      ],
      "correct": 1,
      "explanation": "Crew autonomy makes prod replay and audit painful; wrap with a Flow when shipping."
    }
  ]
}