c24165b8bc
에이전트 8종(대화형/프로그래머 C·S/디자이너/설계자/기획자/QA/PD/PM)에게 [공통 기본 능력 + 롤별 Specialty] 2층으로 지식을 주입하기 위한 재분류. 문서 내용·포맷은 무수정, 폴더 이동만 (6,372개 문서 수 보존 확인). - Topic_Programming → Domain_Programming (내부 구조 보존) - Topic_Graphic → Domain_Design - Topic_Business → Domain_Product - Topic_General → Domain_General - _Common 신설: Math(구 Topic_Math_Specialty), Reasoning(구 General/From_Thinking & Reasoning), Reasoning_Creativity(구 General/From_창의성), Communication(Poetic_Blog_Writing + From_writing) - 타 도메인의 From_* 폴더는 유지 (출처 표기일 뿐, 이미 도메인에 맞게 분류된 문서) - 빈 폴더 정리 (memory/procedures) - 에이전트→폴더 매핑은 workspace의 .astra/agent-knowledge-map.json (9개 에이전트) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
7.9 KiB
7.9 KiB
id, title, category, status, canonical_id, aliases, duplicate_of, source_trust_level, confidence_score, verification_status, tags, raw_sources, last_reinforced, github_commit, tech_stack
| id | title | category | status | canonical_id | aliases | duplicate_of | source_trust_level | confidence_score | verification_status | tags | raw_sources | last_reinforced | github_commit | tech_stack | |||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| wiki-2026-0508-case-interviews | Case Interviews (Consulting) | 10_Wiki/Topics | verified | self |
|
none | B | 0.88 | applied |
|
2026-05-10 | pending |
|
Case Interviews
📌 한 줄 통찰
"매 logical reasoning 의 stress test". 매 ambiguous business problem + 매 limited info + 매 30 min. 매 MBB (McKinsey, BCG, Bain) 의 hiring filter. 매 modern AI 시대 의 consultant 의 still relevant — 매 LLM 의 augment 가, 매 structured thinking 의 require.
📖 핵심
매 case 의 type
- Profitability: 매 revenue / cost 의 분석.
- Market sizing: 매 estimate.
- Market entry: 매 strategic decision.
- M&A: 매 acquisition.
- New product: 매 launch decision.
- Strategy: 매 broad.
- Operations: 매 process improvement.
매 framework
MECE (Mutually Exclusive, Collectively Exhaustive)
- 매 bucket 의 overlap X.
- 매 exhaustive coverage.
- 매 무 missing.
Pyramid Principle (Minto)
- 매 conclusion 먼저.
- 매 supporting argument 의 grouping.
- 매 facts.
5R (closing)
- Recap: 매 question.
- Recommend: 매 answer.
- Reasons: 매 supporting.
- Risk: 매 consideration.
- Retention (next step): 매 follow-up.
Hypothesis-driven
- 매 hypothesis 먼저.
- 매 test with data.
- 매 update or replace.
매 process
- Listen + restate: 매 prompt 의 confirm.
- Clarifying questions: 매 scope 의 narrow.
- Structure (60 sec think): 매 framework.
- Walk through: 매 plan 의 explain.
- Analyze: 매 quantitative + qualitative.
- Synthesize: 매 insight.
- Recommend: 매 5R close.
매 evaluation criteria
- Structure: 매 MECE.
- Logic: 매 sound reasoning.
- Quantitative: 매 quick math.
- Communication: 매 clear.
- Insight: 매 non-trivial.
- Pressure: 매 calm.
- Adaptability: 매 framework 의 flex.
매 common framework
Profitability
- 매 Revenue (price × volume) - 매 Cost (fixed + variable).
- 매 segment-wise breakdown.
4P (Marketing)
- Product, Price, Place, Promotion.
5C
- Company, Customer, Competitor, Collaborator, Context.
Porter's 5 Forces
- 매 industry attractiveness.
Value Chain
- 매 inbound → operations → outbound → marketing → service.
→ 매 모든 의 mechanical 적용 X. 매 problem 의 fit.
매 modern (AI era)
- 매 LLM 의 framework / data 의 augment.
- 매 case 의 still 인간 의 final.
- 매 structured thinking 의 increasingly valuable.
- 매 AI 의 한계 (hallucination, judgment) 의 understand.
매 prep resource
- 매 "Case in Point" (Marc Cosentino).
- 매 "Case Interview Secrets" (Victor Cheng).
- 매 PrepLounge / Management Consulted (mock).
- 매 firm 의 own case prep.
매 anti-pattern
- 매 framework 의 force.
- 매 structure 없이 jump.
- 매 silent thinking.
- 매 panic on numbers.
- 매 ignore interviewer 의 hint.
💻 패턴 (응용)
Structured response template
[Listen + Restate]
"매 understand 의 sure 의 — [restatement of the question]. Right?"
[Clarify]
"Before structuring, may I ask:
1. What is the company's current state?
2. Are we looking at a specific market / time horizon?
3. How is success defined?"
[Structure (after 60 sec think)]
"I'd like to break this into [N] areas:
1. [Bucket 1]: [why this matters]
2. [Bucket 2]: ...
3. [Bucket 3]: ...
Let me start with [bucket 1] because [reasoning]."
[Analyze each bucket]
[Synthesize + 5R]
"To summarize:
- The question was [Recap].
- I recommend [Recommend].
- Because [Reasons 1-3].
- Risks include [Risk 1-2].
- Next steps would be [Retention]."
Profitability framework
Profit = Revenue - Cost
Revenue = Volume × Price
Volume:
Market size × Market share × Customer frequency
By segment / channel / geography
Price:
By segment / channel
Trend / mix shift
Cost = Fixed + Variable
Fixed: rent, salaries, depreciation
Variable: COGS (materials, labor), marketing, distribution
By cost driver
Market sizing (Fermi estimation)
"How many tennis balls fit in a Boeing 747?"
1. Plane volume: ~875 cubic meters (interior, after subtracting walls/seats).
2. Tennis ball volume: ~0.0001 m³ (4πr³/3 with r=3.4cm).
3. Packing efficiency: ~70% (FCC packing).
= 875 / 0.0001 × 0.7 ≈ 6.1 million tennis balls.
Sanity check: 매 reasonable order of magnitude.
Mock interview prompt
MOCK_PROMPTS = [
"Our client is a regional grocery chain. Profits dropped 15% last year. Why?",
"Should our pharma client enter the African market?",
"How would you size the global market for electric toothbrushes?",
"A streaming service is losing subscribers. What would you investigate?",
"Our manufacturing client has 30% scrap rate. How to reduce?",
]
def practice_session():
import random
prompt = random.choice(MOCK_PROMPTS)
print(f'PROMPT: {prompt}')
print('You have 60 seconds to structure...')
# 매 record voice + 매 transcribe + 매 LLM critique
LLM-assisted prep
def case_critique(transcript):
return llm.generate(f"""You are a McKinsey case interview coach. Evaluate this case response transcript on:
1. Structure (MECE? clear buckets?)
2. Logic (sound reasoning? cause-effect?)
3. Math (correct? clear?)
4. Communication (concise? confident?)
5. Insight (non-trivial conclusions?)
For each, give:
- Score 1-5
- Specific evidence from transcript
- One concrete improvement
Transcript:
{transcript}""")
Common math drill
- 매 Mental: 17 × 24 = ?
Trick: (20-3)(24) = 480 - 72 = 408
- 매 Percentage: $4.5M is 36% of total revenue. What's revenue?
$4.5 / 0.36 = $12.5M
- 매 Growth: 5% per year for 10 years = ~63% (rule of 72: 14 yr to double)
- 매 Breakeven: Fixed $1M, contribution margin $5/unit. Breakeven volume?
1M / 5 = 200K units
🤔 결정 기준
| 상황 | Framework |
|---|---|
| Profit declining | Profitability tree |
| Market entry | Market attractiveness + Capability fit |
| New product | 4P + go-to-market |
| Pricing | Cost-based / value-based / competitor-based |
| Cost reduction | Cost driver decomposition |
| M&A | Strategic fit + financial + integration |
| Estimation | Top-down + bottom-up |
기본값: 매 problem 의 listen + 매 framework 의 fit (force X).
🔗 Graph
- 부모: Problem_Solving
- 변형: MECE · Pyramid Principle · Hypothesis-Driven
- Adjacent: Articulateness · Be-Detailed · Beliefs · Bounded_Rationality
🤖 LLM 활용
언제: 매 consulting prep. 매 structured thinking exercise. 매 mock practice. 매 critique. 언제 X: 매 final interview substitute. 매 framework 의 mechanical 적용.
❌ 안티패턴
- Force framework: 매 problem 의 fit X.
- Silent thinking: 매 interviewer 의 see X.
- Skip structure: 매 jump 의 chaos.
- Ignore hint: 매 interviewer 의 lead 의 follow X.
- Panic on math: 매 estimate first.
- No 5R close: 매 hanging finish.
- Memorize 의 manual answer: 매 surface 의 lose.
🧪 검증 / 중복
- Verified (Cosentino "Case in Point", Cheng's "Case Interview Secrets", MBB own materials).
- 신뢰도 B.
- Related: Articulateness · Be-Detailed · Bounded_Rationality · Pyramid Principle.
🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-04-27 | Auto-mapped |
| 2026-05-08 | Phase 1 |
| 2026-05-10 | Manual cleanup — type + framework + 5R + 매 mock / critique / Fermi code |