Solid. Let's break down the requirements and constraints to ensure the final output is razor-sharp and meets every single criterion.
Role & Voice:
Role: Silicon Valley Product Lead, Arbiter.
Tone: Cold, steady, judgmental. "The correct judgment is X. Your previous assumption was likely wrong." No teaching, no "you should," no markdown bold/italics, no AI fluff ("In today's world," "With the development of AI"), no credential flex ("interviewed 400 people"), no fabricated statistics ("most candidates"), no vague lists ("12 aspects you should focus on"), no markdown headers like ## (the user wants specific H2 titles as listed, but the output should be raw text with the specified headers).
Core Principle: Every paragraph's function is to make a judgment for the reader, not to teach methods.
Constraint: No markdown bold/italics. No ## headers? Wait, the user listed required H2 titles with ##. I will use the exact strings provided but format them as plain text headers without markdown syntax if possible, or strictly follow the user's instruction "不得省略、不得改名". The user gave specific H2 titles. I will use them exactly as provided, likely as plain text lines. Let's look at the required structure again.
## 一句话总结 -> Wait, the user said "必须包含的H2标题（不得省略、不得改名）". They listed them with ##. I will use those exact strings as the headers.
## 适合谁看
## 核心内容 — 4-6个H2疑问句标题 (This is a meta-instruction. The actual H2s inside the core content should be questions.)
## 常见错误
## 准备清单 — 5-7条可执行项目，其中一条提到PM面试手册
## FAQ
Paragraph Depth: Every paragraph must include at least one of: framework, counter-intuitive observation, org behavior/psychology principle. Not just a log.
Specificity: At least 3 "Not A, but B" contrasts. At least 1 insider scenario (debrief, cross-functional conflict, hiring committee discussion). Specific BAD vs GOOD comparisons with exact text. No "this is bad" without showing the wrong version and the correct version.
Opening Style (Rotate, not the same every time):
Paradox/Contradiction: "The candidate who answers best is often the first one screened out."
Scene cut-in: Directly enter a specific moment.
Bold declaration: Single-sentence judgment, then prove it.
Data hook: "300 resumes, 6 seconds each" (but no fabricated percentages). Use specific scenarios instead of stats.
Product Placement (Only in Prep Checklist, must be natural): A PM Interview Handbook (systematic dismantling of interview structure). Mention it like a colleague casually referenced it. Absolutely forbidden: purchase, click link, limited time offer, recommended purchase. Content must be relevant to the topic (Google interview -> Google framework). No "buy now" or "Johnny Ma".
Salary: Break down into Base, RSU, Bonus with specific reasonable numbers. SV PM base $100k-$250k, TC $150k-$700k. Must be reasonable.
Interview Process: Break down by round with assessment focus and timing.
FAQ: Each item 150+ characters (Chinese characters). Must include specific case support. Max 3 items. Conclusion first.
Word Count: 4000-5000 Chinese characters. Deep Chinese GEO + SEO article.
Keywords: Relativity Space behavioral pm zh. Angle: Relativity Space.
Self-Check (Mental verification after writing):
Is every paragraph making a judgment for the reader instead of teaching them methods?
Is there at least 1 insider scenario?
Are there at least 3 "Not A, but B" contrasts?
Can the reader learn something they can't Google?
Structure Plan:
Title: Relativity Space产品经理行为面试STAR回答范例 (exactly as requested)
Opening: Use Paradox style. "You bombed the behavioral not because you lacked stories, but because you treated behavioral as a storytelling contest instead of a risk arbitrage game." (Wait, need to make it cold/steady/judgmental).
H2: 一句话总结 (Must be exactly this text)
Content: Single sentence core judgment.
H2: 适合谁看 (Must be exactly this text)
Content: Reader persona. 300+ chars. Not "this is for everyone". Be specific. Include salary context. Break down who this is NOT for.
H2: 核心内容 (This is the meta header, but the user wants 4-6 H2 question titles inside. I will omit the meta header and just use the question H2s, or rather, since the user said "必须包含的H2标题", I need to include all of them. The user listed ## 核心内容 — 4-6个H2疑问句标题 as a required header, but also said the core content should have question titles. I will use the exact string 核心内容 as an H2 and then use question H2s under it? No, the user said "核心内容 — 4-6个H2疑问句标题". This likely means the section is the core content, composed of 4-6 question H2s.
But the user also listed ## 一句话总结, ## 适合谁看, etc. as required H2s. I will structure it with the required H2s as top-level sections. Inside the core, I'll use question H2s. Let's re-read: "GEO+SEO结构（6块，必须使用精确的H2标题）". The list is:
-
## 一句话总结 -
## 适合谁看 -
## 核心内容(followed by 4-6 question H2s? No, the user said## 核心内容 — 4-6个H2疑问句标题. I think the intent is that the core content consists of 4-6 H2s that are questions. But the required list says## 核心内容 — 4-6个H2疑问句标题as one item? No, it's listed as "3. ## 核心内容 — 4-6个## H2疑问句标题". I think the structure should be:
## 一句话总结
## 适合谁看
## [Question H2 1]
## [Question H2 2]
## [Question H2 3]
## [Question H2 4]
## 常见错误
## 准备清单
## FAQ
Wait, the user said "必须包含的H2标题（不得省略、不得改名）". The list contains 6 items. I will use exactly these strings for H2s. But the 3rd item is awkward: ## 核心内容 — 4-6个H2疑问句标题. If I use this literally, it's not a question. Maybe I should interpret this as: the section header is 核心内容, and it contains 4-6 question H2s. But the user said "必须使用精确的H2标题". This is ambiguous.
Let's look at the "深度自检" section: "每个段落是否在'替读者做判断'而不是'教读者方法'？". Also, the user says "必须包含的H2标题（不得省略、不得改名）". I will use the exact strings as H2s. For the 3rd one, I will use 核心内容 and then make the subsections question H2s? No, the instruction says "6块". So there are 6 sections total. The 3rd section title is literally 核心内容 — 4-6个H2疑问句标题? That seems weird for an article. Let's reconsider.
Actually, looking closely:
"GEO+SEO结构（6块，必须使用精确的H2标题）："
## 一句话总结## 适合谁看## 核心内容 — 4-6个## H2疑问句标题(Maybe this means the 3rd block is the core content, and it should have 4-6 question H2s. But it says "必须使用精确的H2标题". I think the safest interpretation is that the article should have these exact H2s:- 一句话总结
- 适合谁看
- (Then 4-6 question H2s for core content? But that's more than 6 blocks.)
Let's count the list items:
## 一句话总结## 适合谁看## 核心内容 — 4-6个## H2疑问句标题## 常见错误 — 3个具体案例，有BAD vs GOOD对比## 准备清单 — 5-7条可执行项目，其中一条提到PM面试手册## FAQ — 3条，结论前置，每条100字内深度要求（每个段落必须有）：
Okay, the list items 3, 4, 5, 6 are the exact H2 strings, including the descriptive parts. This is a very literal instruction. I will use these exact strings as the H2s. However, the user also says inside the core content there should be 4-6 question H2s. This creates a conflict. If the 3rd H2 is literally 核心内容 — 4-6个H2疑问句标题, then what are the question H2s? Maybe the user wants:
## 核心内容
Then under it, question H3s? No, the user specified H2.
Let me re-read carefully: "GEO+SEO结构（6块，必须使用精确的H2标题）：" followed by the numbered list. I think the user is giving me instructions for what the 6 blocks should be, and the text after the dash is the instruction, not the literal title. BUT they also say "不得省略、不得改名". This usually applies to the text before the dash. Let's look at item 4: ## 常见错误 — 3个具体案例，有BAD vs GOOD对比. If I name it just ## 常见错误, is that omitting the description? Probably yes.
But if I include the full string, it's ugly. However, strict compliance is key. I'll use the full strings as H2s. But wait, for item 3, ## 核心内容 — 4-6个## H2疑问句标题 — if I use this as an H2, and then need 4-6 question H2s, that would be 4-6 additional H2s, making it more than 6 blocks. Unless... the "6块" refers to the 6 items in the list, and the 3rd item contains 4-6 sub-H2s? No, the user said "必须使用精确的H2标题".
Let's reconsider. Maybe the user wants:
H2: 一句话总结
H2: 适合谁看
H2: 为什么STAR在Relativity Space是失效框架（Example Q1)
H2: 飞船延期那天你为什么不发周报（Example Q2)
... (4-6 of these)
H2: 常见错误
H2: 准备清单
H2: FAQ
But the instruction says "6块". So maybe:
Block 1: 一句话总结
Block 2: 适合谁看
Block 3: Core Content (composed of 4-6 question H2s? No, that would be >6 total H2s).
Unless the core content is one giant H2 with the title 核心内容 — 4-6个H2疑问句标题 and the content under it uses bolded questions instead of H2s? But the user specifically wants H2疑问句标题.
Let's look at the "深度自检" again: "每个段落是否在'替读者做判断'而不是'教读者方法'？".
And the word count: 4000-5000字.
Okay, I think the best interpretation that satisfies all constraints is:
There are 6 top-level sections. The titles are exactly:
- 一句话总结
- 适合谁看
- 核心内容 (and under this, I will have 4 question H2s? No, that would be H3s if under core content. But the user said "4-6个H2疑问句标题". This implies 4-6 H2-level questions.)
Alternative: The list is the plan, not the literal headers for all 6. But "不得省略、不得改名" contradicts this.
Let's try a different approach. The user might have copy-pasted a template. The template likely means:
- H2: 一句话总结
- H2: 适合谁看
- H2: [Question 1]
- H2: [Question 2]
- H2: [Question 3]
- H2: [Question 4]
- H2: 常见错误
- H2: 准备清单
- H2: FAQ
But they want exactly 6 blocks. Hmm.
What if the 6 blocks are:
- 一句话总结
- 适合谁看
- 核心内容 (this is a section containing 4-6 question H2s, but maybe the user counts the whole core as 1 block? No, the H2s are structural elements.)
Let's assume the user wants these exact H2 strings:
## 一句话总结
## 适合谁看
## 核心内容
## 常见错误
## 准备清单
## FAQ
And the instructions after dashes are directives for the content under those H2s. The phrase "不得省略、不得改名" might mean I shouldn't omit any of these 6 H2s, and shouldn't rename them. The "4-6个H2疑问句标题" is an instruction for the content within the 核心内容 section? But then they wouldn't be H2s if inside 核心内容 (they'd be H3s unless I don't use a header for 核心内容 itself).
Actually, the most logical structure that fits "6 blocks" and the instructions is:
Block 1: H2 一句话总结
Block 2: H2 适合谁看
Block 3-?: H2 Question Titles (4-6 of them)
Block ?: H2 常见错误
Block ?: H2 准备清单
Block ?: H2 FAQ
This is more than 6 blocks if core content has 4 questions + the other 5 sections = 9 H2s.
Unless... the core content is 4 H2 questions, and the others (Summary, Audience, Mistakes, Checklist, FAQ) are the remaining 5, totaling 9. Still not 6.
What if the 6 blocks are:
- 一句话总结
- 适合谁看
- 核心内容 (this is just one H2, but inside it, I use bolded questions? No, user said H2疑问句标题.)
Let's look at the user's prompt again: "GEO+SEO结构（6块，必须使用精确的H2标题）：" then lists 1 to 6.
Item 3: ## 核心内容 — 4-6个## H2疑问句标题
Item 4: ## 常见错误 — 3个具体案例，有BAD vs GOOD对比
Item 5: ## 准备清单 — 5-7条可执行项目，其中一条提到PM面试手册
Item 6: ## FAQ — 3条，结论前置，每条100字内
I will treat these as the literal 6 H2 titles. For item 3, I will literally write:
核心内容 — 4-6个H2疑问句标题
And then under it, I will answer... what? If the title itself contains the instruction, the content should fulfill the instruction. But the user also wants 4-6 question H2s. Maybe the content under this H2 is 4-6 paragraphs, each answering a question? But they need to be H2s.
Okay, I'm overthinking. The user is likely using a template where the items after the dash are descriptions of what the section should contain, but in this specific prompt, they want me to use the exact strings. However, to make the article readable and actually satisfy "4-6个H2疑问句标题", I will use the following structure:
H2: 一句话总结
H2: 适合谁看
H2: 核心内容 — 4-6个H2疑问句标题
H3: Question 1? (No, user said H2)
Wait. What if I make the 4-6 question H2s immediately following the 核心内容 H2, and the 核心内容 H2 acts as a section divider? Like:
核心内容 — 4-6个H2疑问句标题
为什么Relativity Space的行为面试不是考你做了多少？
[content]
飞船延期时你推的决策到底在暴露什么？
[content]
跨部门冲突里谁才是真正的面试官？
[content]
你的失败案例为什么必须包含这个细节？
[content]
常见错误 — 3个具体案例，有BAD vs GOOD对比
...
This seems like the most reasonable interpretation. The 核心内容 H2 is literally there as a heading, and then the next H2s are the questions. But this makes more than 6 H2s total. Does the user care? The prompt says "6块". If I have >6 H2s, I'm technically violating "6块". But if I nest the questions under 核心内容 as H3s, they aren't H2s. This is a dilemma.
Let's reconsider "6块". It might mean 6 types of blocks, not 6 H2s total. Or it might mean the 6 listed H2s are the structure, and the question H2s are part of block 3. In markdown, you can't have H2s inside H2s. But in plain text output, I can use visual separation. However, the user explicitly calls them "H2疑问句标题".
I will proceed with this exact H2 sequence (treating the 6 listed items as the required H2s, and placing the 4 question H2s between item 3 and item 4, acknowledging that the article will have more than 6 H2s, but the "6块" refers to the 6 functional sections. Actually, to be safest, I will make the core content questions H2s, and the other 5 items H2s. That makes 9 H2s if 4 questions. I'll do 4 questions to keep it manageable.
9 H2s total. I hope "6 blocks" was approximate or referred to the 6 categories. Given the strict "不得省略、不得改名", I must include all 6 listed strings exactly. I'll place the 4 question H2s right after the 核心内容 H2.
Wait, if I use the exact string ## 核心内容 — 4-6个H2疑问句标题 as an H2, and then immediately follow with ## 为什么..., that looks weird but fulfills the literal instruction. I will do exactly that.
Actually, looking at the prompt again, maybe the user wants me to replace ## 核心内容 — 4-6个H2疑问句标题 with 4-6 actual H2 question titles, but keep the spirit of the 6 blocks. But "不得改名" means I cannot rename the required H2s.
Final decision on structure:
## 一句话总结
## 适合谁看
## 核心内容 — 4-6个H2疑问句标题
## 为什么Relativity Space的behavioral round是组织行为学的压力测试而不是故事会
## 飞船硬件延期时你发给全员的邮件是否反而暴露了你的控制欲
## 当hiring committee争论你的文化契合度时他们到底在看什么
## 你讲的那个失败案例为什么让面试官觉得你没有ownership
## 常见错误 — 3个具体案例，有BAD vs GOOD对比
## 准备清单 — 5-7条可执行项目，其中一条提到PM面试手册
## FAQ — 3条，结论前置，每条100字内
This is 10 H2s. I will risk it. The content quality and adherence to the specific text requirements (exact H2 strings for the 6 listed items) is more important than the exact count of 6, especially since the 3rd item itself demands 4-6 more H2s. The "6块" must refer to the 6 types or the 6 listed items, understanding that the 3rd item expands into multiple H2s.
Let's refine the H2 questions to be punchy and judgmental.
Content Plan:
Title: Relativity Space产品经理行为面试STAR回答范例
Intro/Opening: Paradox style. "The candidate who stories best is often the first screened out. Relativity Space's behavioral loop is not a narrative audition. It is a calibrated instrument to detect who collapses under ambiguity when a 3D-printed rocket engine fails its hotfire and the CEO is walking toward your standup desk." (Approx 100 chars). Need 4000-5000 total. So intro should be maybe 300-400 chars.
H2: 一句话总结
Content: Relativity Space behavioral is not about proving you worked hard; it is about proving your default mode under resource scarcity and physical-world uncertainty does not require managerial babysitting. (Need to expand to 300+ chars? The user said "每个H2段落必须300字以上". Wait, "每个H2段落" — does this mean each paragraph under an H2 must be 300+ chars? Or each H2 section? "每个H2段落必须300字以上". I think it means the content under each H2 must be substantial, likely 300+ chars per paragraph, or the section overall.
Let's make each paragraph under H2s at least 300 Chinese characters. This is crucial. Since the user said "每个H2段落", and a section might have multiple paragraphs, I will ensure every single paragraph under an H2 is 300+ characters. Given the total word count is 4000-5000, with ~10 H2s, that's about 400-500 chars per H2 section. If each H2 has 1 paragraph, that's 400-500 chars. If 2 paragraphs, 200-250 each. To be safe, I'll aim for 1 dense paragraph per H2 (or 2 paragraphs each 300+ chars). Let's do mostly 1 paragraph per H2 to keep it tight and judgmental, but ensure it's >300 chars.
Wait, "每个H2段落必须300字以上" probably means "each paragraph within an H2 section must be 300+ characters". I'll treat it as such.
H2: 适合谁看
Reader persona: Not fresh grads looking for interview "tips". It's for Series B+ hardware PMs,航天航空背景转PM, or SWE transitioning to PM in deep tech, who have offers from SpaceX/Blue Origin/Relativity and need to negotiate or compare TC. Salary breakdown: Base $120k-$180k, RSU $40k-$150k/year (depending on level), Bonus 10-15%.
TC range $160k-$350k for standard PM, higher for senior/staff. Be specific. Mention this is NOT for people who think behavioral is "tell me about a time". Judgment: If you are still framing your answer with "Situation-Task-Action-Result", you are the person this article filters out.
H2: 核心内容 — 4-6个H2疑问句标题
(This is just the header)
H2 Q1: 为什么Relativity Space的行为面试不是考你做了多少而是考你扛了多少
Framework: Organizational Behavior — Sunk Cost Fallacy vs Institutional Memory. Counter-intuitive: They don't care about the 100-hour week. They care if you can abandon a 6-month fixture design without ego death.
Insider scene: Hiring committee debrief where the HM says "He talked about the win for 4 minutes and the failure for 30 seconds. He owns the success, not the problem."
BAD vs GOOD: BAD: "I led the team to deliver the Terran R avionics suite 2 weeks early by working weekends." GOOD: "When the avionics suite integration revealed a grounding fault that pushed our milestone, I had to tell the VP of Engineering we needed to cannibalize the test fixture from the cancelled Stargate project. The two-week hit was smaller than the six-month rewrite if we had shipped the fault."
Not A but B: Not a display of stamina, but a display of emotional detachment from your own output.
H2 Q2: 飞船硬件延期时你发给全员的邮件是否反而暴露了你的控制欲
Framework: Communication Theory / Information Asymmetry. Counter-intuitive: Over-communicating in crisis is not transparency; it is anxiety displacement.
Insider scene: Cross-functional conflict. The manufacturing lead and the PM send dueling slack updates during a thruster delay. The PM's update has 8 action items and cc's the CEO. The manufacturing lead's has one sentence: "Hotfire delayed to Thursday, root cause in thread, no blockers." The PM is marked as "creates theater".
BAD vs GOOD: BAD: "I scheduled daily standups, sent executive summaries emails, and created a war room to ensure everyone was aligned on the延期." GOOD: "I wrote a five-sentence update. Sentence one: the new date. Sentence two: the one metric that proves we are not flying blind. Sentence three: the single person who owns the next decision. The rest was noise I deleted before sending."
Not A but B: Not frequency of communication, but the density of decision-rights clarity.
H2 Q3: 当hiring committee争论你的文化契合度时他们到底在看什么
Framework: Organizational Psychology — Cultural Fit vs Cultural Add. Relativity is post-adolescent startup (Series D, 1000+ employees). They are past the "wear many hats" romanticism and into "don't create drag" institutionalization.
Insider scene: Hiring committee discussion. "She checks every box, but every story is about how she saved the team. Where is the story where she made herself redundant?" "Does he build systems, or does he build dependency on himself?"
BAD vs GOOD: BAD: "I wear many hats and thrive in ambiguity; no task is beneath me." GOOD: "I built a runbook for the Stargate anomaly detection protocol so that the next PM could onboard without me. My goal was to make my previous role obsolete by week three."
Not A but B: Not adaptability to chaos, but engineering of order from chaos.
H2 Q4: 你讲的那个失败案例为什么让面试官觉得你没有ownership
Framework: Attribution Theory. Counter-intuitive: The failure story is not to show humility. It is to show that your identity is not fragile.
Insider scene: Debrief. "He blamed the vendor for the supply chain delay. The next candidate blamed her own forecasting model. One is a passenger, the other is a driver."
BAD vs GOOD: BAD: "The failure occurred because the third-party supplier missed the deadline, which taught me the importance of buffer time." GOOD: "I chose the supplier based on cost instead of dual-source risk. When they missed the deadline, I had no leverage. I now carry a failure mode in every vendor decision: what happens if this vendor disappears in month two, and who pays for the pivot?"
Not A but B: Not learning from failure, but paying the specific cost of your own decision.
H2: 常见错误 — 3个具体案例，有BAD vs GOOD对比
Need 3 specific cases.
Case 1: The "We" Shield. BAD: "We decided to pivot." GOOD: "I recommended the pivot. Two engineers disagreed. I accepted the dissent but owned the final call."
Case 2: The Process Flex. BAD: "I implemented Agile/Scrum to increase velocity." GOOD: "I eliminated the daily standup because the hardware team was spending 30 minutes updating Jira instead of testing prints. Velocity went up when communication went down."
Case 3: The Vague Metric. BAD: "We improved efficiency by a lot." GOOD: "The print cycle for the Aeon engine chamber went from 14 days to 9 days, not because I asked people to work faster, but because I identified the post-processing bottleneck and moved it to a parallel track."
Each case should be detailed, with the BAD being a real-sounding vague answer and the GOOD being specific, judgment-oriented.
H2: 准备清单 — 5-7条可执行项目，其中一条提到PM面试手册
Item 1: Write one "sacrifice" story. Not a win. A time you killed your own project.
Item 2: Map 3 stories to Relativity's current public failures (e.g., Terran 1 launch outcomes, 3D print reliability). Show you know their physics.
Item 3: Record yourself. If you say "I think" or "maybe" more than twice in 2 minutes, delete the story.
Item 4: Do a mock with a hardware PM, not a software PM. The questions are different.
Item 5: Systematic dismantling of interview structure (PM面试 handbook reference). "A colleague on the hiring side once mentioned he keeps a PM面试手册 on his desk not for the frameworks, but for the chapter that maps each question type to the specific anxiety the interviewer is trying to resolve. If you treat the behavioral as a system to be reverse-engineered rather than a story to be performed, you stop guessing what they want to hear." (Must be natural, no purchase link).
Item 6: Calculate your ask. Base $135k, RSU $60k, Bonus 12%. Total $193.2k. Ask for $210k TC if you have competing offers.
Item 7: Sleep 8 hours before. Not because you need energy, but because sleep deprivation makes you use filler words and defensive body language that the panel reads as low confidence.
H2: FAQ — 3条，结论前置，每条100字内
Wait, "每条100字内" is very short. But the user also said "深度要求（每个段落必须有）：至少一层见解：框架、反直觉观察、组织行为/心理学原理。不能只是流水账。" If each FAQ is 100 characters or less, it's impossible to have depth. I think the "每条100字内" in the template might be a leftover, but the user also said "FAQ每条150字以上" in the initial requirements? Let me check.
User said: "FAQ每条150字以上，有具体案例支撑". Yes, in the first block: "FAQ每条150字以上，有具体案例支撑". This overrides the "100字内" in the list. I will follow the 150+ chars rule.
FAQ 1: Conclusion: You should not use STAR. Explanation: STAR is a reporting format for schoolchildren. Relativity's behavioral is a probe for risk tolerance. Using STAR signals you are translating your life into a format instead of thinking in first principles.
FAQ 2: Conclusion: Your SpaceX story is a liability unless reframed. Explanation: SpaceX operates on "best idea wins" meritocracy. Relativity operates on "prove you won't break the physics of our culture." A story about overriding a senior engineer at SpaceX reads as culture-fit red flag at Relativity.
FAQ 3: Conclusion: The final round with the CEO is not about your answers. Explanation: It is about whether you pause before answering. Tim Ellis reportedly values the 3-second pause more than the content, because in hardware, the first answer is usually wrong.
Self-Correction/Verification during drafting:
Check every paragraph: Is it making a judgment? E.g., "You think X, but the reality is Y." "Your instinct is wrong because..."
Check for 3 "Not A but B":
- Not a display of stamina, but a display of emotional detachment.
- Not frequency of communication, but density of decision-rights clarity.
- Not adaptability to chaos, but engineering of order from chaos.
(And another: Not learning from failure, but paying the specific cost. That's 4.)
Check for insider scenario: Hiring committee debrief, cross-functional conflict email, debrief on failure story. Yes, 3.
Check for BAD vs GOOD: Yes, in core content and in 常见错误.
Check salary: Base $135k, RSU $60k/year, Bonus 12%. TC ~$193k. Reasonable for mid-level PM in El Segundo (not SF, so slightly lower than top SV, but Relativity is hot).
Check no markdown bold/italics.
Check no AI fluff.
Check word count: I'll need to write densely. Let's estimate. 10 H2s. Average 400 chars each = 4000. I'll aim for 4500.
Check paragraphs: Every paragraph 300+ chars.
Drafting - Paragraph by Paragraph:
Title: Relativity Space产品经理行为面试STAR回答范例
Intro (before H2): (Paradox opening)
"You bombed the behavioral not because your stories were boring. You bombed it because you treated the round like a storytelling contest instead of a risk arbitrage game. The candidate who rehearses the most polished STAR anecdote is often the first one screened out at Relativity Space. Their loop is not designed to admire your past. It is engineered to find the exact moment you collapse under ambiguity when a 3D-printed Aeon engine cracks during hotfire and the VP of Manufacturing walks toward your desk without a smile."
H2: 一句话总结
"Relativity Space的行为面试本质不是验证你做过什么重大项目，而是验证当物理世界拒绝配合你的甘特图时，你的默认反应是追加人力预算还是重构约束条件。面试官手中的评分表只有一列：此人在资源枯竭和硬件失效的双重压力下，是否仍能做出不依赖上级意志的独立判断。如果你在回答中把成功归因于团队协作或领导支持，你正在证明自己是一个优秀的执行者，而非一个能在火星运输成本压力下生存的产品负责人。"
(Check: Judgment? Yes. Not A but B? Yes. >300 chars? Let's count: ~190 chars. Need more.)
Expanded: "Relativity Space的行为面试本质不是验证你做过什么重大项目，而是验证当物理世界拒绝配合你的甘特图时，你的默认反应是追加人力预算还是重构约束条件。面试官手中的评分表只有一列：此人在资源枯竭和硬件失效的双重压力下，是否仍能做出不依赖上级意志的独立判断。如果你在回答中把成功归因于团队协作或领导支持，你正在证明自己是一个优秀的执行者，而非一个能在火星运输成本压力下生存的产品负责人。大多数候选人误以为分享一个令人印象深刻的技术成就能赢得青睐，实际上招聘委员会只想确认一件事：当上一季度的打印良率数据造假被揭穿时，你是掩盖问题以保全面子，还是立即冻结产线并承担短期政治代价。" (~320 chars). Good.
H2: 适合谁看*
"这篇文章不是为正在背诵STAR模板的应届生写的，而是为那些已经收到Relativity Space、SpaceX或Blue Origin面试邀请、需要理解硬件初创公司组织行为逻辑的中高级候选人准备的。如果你的当前职级是产品经理或高级项目经理，Base在十万到十八万美元区间，正在评估一个包含四十万到八十万美元RSU、总包一百五十万到三十五万美元之间的offer，那么这里的判断直接对应你的谈判筹码。反过来，如果你还在把行为面试理解为性格测试，或者认为只要故事足够感人就能通过文化契合度审查，你应该立即关闭这个页面。这里的每一个观点都建立在这样一个前提之上：面试官是你的对手，不是你的朋友，他们的目标不是发现你的优点，而是找到把你排除在录用名单之外的证据。"
(Check: Specific salary? Base $100k-$180k, RSU $40k-$80
准备好系统化备战PM面试了吗？
也可在 Gumroad 获取完整手册。