01核心结论Key takeaways
以下数字为 Metix AI 数据库口径（数据时点约 2026 年上半年），统计对象 = 当前在职于 OpenAI 的技术人才（工程与研究序列）。所有数字均为聚合统计，报告不展示任何个人信息。The figures below reflect the scope of the Metix AI database (data as of roughly the first half of 2026). The population = technical staff currently employed at OpenAI (engineering and research tracks). All numbers are aggregate statistics; the report displays no personal information.
中位职业经验 13 年，博士仅 14.4%，技能榜由 Python、分布式系统、Java、C++ 领衔。OpenAI 买的是已经把大系统跑起来、并扛过规模的人，而不是论文产出者。Median career experience is 13 years, only 14.4% hold a PhD, and the skill leaderboard is led by Python, distributed systems, Java, and C++. OpenAI is buying people who have already stood up large systems and carried them through scale — not paper producers.
中位在职仅 17 个月，35% 是过去一年内入职，近三分之二在两年内加入。OpenAI 在 2026 年本质上是一家顶着大牌的两岁创业公司——也意味着大量人才正处在最易被打动的窗口。Median tenure is just 17 months, 35% joined within the past year, and nearly two-thirds joined within two years. In 2026 OpenAI is essentially a two-year-old startup wearing a marquee name — which also means a large share of its talent sits in the most persuadable window.
702 人此前在 Google（占近四分之一），其后是 Meta、微软、苹果、亚马逊。OpenAI 主要从大厂基础设施团队取水，而非在 AI 实验室之间互挖。702 people previously worked at Google (nearly a quarter), followed by Meta, Microsoft, Apple, and Amazon. OpenAI draws mainly from Big Tech infrastructure teams, rather than poaching across AI labs.
57% 共用同一个头衔 Member of Technical Staff——20 年老兵、连续创业者、应届博士共用一行 title。想评估 OpenAI 的人，不能看头衔，只能看他真正做过什么。57% share a single title — Member of Technical Staff — with 20-year veterans, serial founders, and fresh PhDs all on the same line. To assess anyone at OpenAI, you can't read the title; you have to read what they've actually built.
02近三分之二的人，过去两年才加入Nearly two-thirds joined only in the past two years
任期是判断一家公司扩张速度最直接的信号。OpenAI 现任技术人才的中位在职仅Tenure is the most direct signal of how fast a company is scaling. The median tenure of OpenAI's current technical staff is just 17 个月17 months，超过三分之一入职不到一年。换句话说，这是一支高速换血、仍在快速搭建的队伍。, with more than a third on board for under a year. In other words, this is a team turning over fast and still building at speed.
03一支资深团队：70% 有 10 年以上经验A senior team: 70% have 10+ years of experience
组织虽新，人却不新。中位职业经验The org is new, but the people aren't. Median career experience is 13 年13 years，应届与初级是例外而非主体。OpenAI 在做的，是把资深经验高度集中，再压进一套刻意扁平的职级里（见第 07 节）。, and new grads and juniors are the exception, not the core. What OpenAI is doing is concentrating senior experience heavily, then compressing it into a deliberately flat leveling system (see Section 07).
04技能榜是工程，不是论文The skill leaderboard is engineering, not papers
把 OpenAI 技术人才的硬技能排个序，榜首是系统与工程语言——Python、分布式系统、Java、C++；深度学习、NLP 这些「研究味」标签排在它们下面，而不是上面。这是一支为「把大系统跑在生产环境里」而搭的队伍。Rank the hard skills of OpenAI's technical staff and the top is systems and engineering languages — Python, distributed systems, Java, C++; the more research-flavored tags like deep learning and NLP sit below them, not above. This is a team built to run large systems in production.
05头号进水管是 Google——大厂，而非实验室The number-one feeder is Google — Big Tech, not a lab
把每个人的过往雇主摊开看，Lay out everyone's previous employers and Google 高居榜首：702 名现任 OpenAI 员工此前在 Google，sits firmly on top: 702 current OpenAI employees previously worked at Google, 接近四分之一nearly a quarter，比微软和苹果加起来还多；其后是 Meta（573）与微软（344）。OpenAI 主要从大厂基础设施团队取水，远多于在 AI 实验室之间互挖。榜单上唯一的非大厂名字 Statsig，来自一次收购，而非常规挖角。 — more than Microsoft and Apple combined; followed by Meta (573) and Microsoft (344). OpenAI draws mainly from Big Tech infrastructure teams, far more than it poaches between AI labs. The one non-Big-Tech name on the list, Statsig, comes from an acquisition rather than routine recruiting.
06斯坦福、伯克利、MIT，以及一条很深的中国管道Stanford, Berkeley, MIT — and a deep China pipeline
学校分布是熟悉的顶尖 CS 名单，由The school distribution is the familiar top-CS lineup, led by 斯坦福Stanford 与 and 伯克利Berkeley 领衔。少被提到的是一条很深的中国管道——清华、北大、上海交大都排得很靠前，通常是赴美读研之前的本科一站。. What gets less attention is a deep China pipeline — Tsinghua, Peking University, and Shanghai Jiao Tong all rank high, typically as the undergraduate stop before grad school in the US.
长尾中的其他中国院校：Other Chinese schools in the long tail: 北大 · 35Peking University · 35上海交大 · 34Shanghai Jiao Tong · 34——美国研究生管道里相当一部分人的本科来源。 — the undergraduate origin of a sizable share of the US grad-school pipeline.
07一个头衔统治一切：Member of Technical StaffOne title rules them all: Member of Technical Staff
OpenAI 用一套极度扁平的职级。OpenAI runs an extremely flat leveling system. 57% 的技术人才顶着某种形式的「Member of Technical Staff」——20 年的分布式系统老兵、出走的创业者、刚毕业的博士，共用同一行 title。内部当然有层级，但从外部看，这张组织图是刻意「不可读」的。 of technical staff carry some form of “Member of Technical Staff” — a 20-year distributed-systems veteran, a founder who walked away from a startup, and a freshly minted PhD all share the same line. There are internal levels, of course, but from the outside this org chart is deliberately unreadable.
其余头衔的长尾The long tail of other titles
Software Engineer · 52Researcher · 45Solutions Engineer · 41Research Scientist · 38Research Engineer · 29Member of Data Science Staff · 28Applied AI · 27Forward Deployed Engineer · 26Solutions Architect · 24除 MTS 之外，全是又长又薄的尾巴。Beyond MTS, it's all a long, thin tail.
08这份透视怎么用How to use this X-Ray
如果你在和 OpenAI 抢人If you're competing with OpenAI for talent
主攻入职不足 24 个月的那一层（占全员近三分之二，仍在股权兑现窗口内）；从 Google、Meta、微软这些大厂基础设施团队取水，而不是只盯实验室；别用头衔筛人——57% 看起来都一样，要读履历底下真正做过的系统。Target the layer with under 24 months of tenure (nearly two-thirds of all staff, still inside the equity vesting window); draw from Big Tech infrastructure teams like Google, Meta, and Microsoft rather than fixating on the labs; and don't screen by title — 57% look identical, so read the actual systems built underneath the resume.
如果你想加入这类团队If you want to join a team like this
拿出你真正构建并扛过规模的系统——这里的中位线是 13 年的工程交付，而不是引用数。没有博士也没关系，你属于 86% 的多数。早期人才的例外通道存在，但要靠顶级实习、竞赛名次或已发表的成果来过线。Show the systems you've actually built and carried through scale — the median bar here is 13 years of engineering delivery, not citation count. No PhD is fine; you'd be part of the 86% majority. An exception lane for early-career talent exists, but you clear it with top-tier internships, competition rankings, or published results.
这只是一家公司。任何一家，我们都能透视。This is just one company. We can X-Ray any of them.
本报告由 Metix AI 的人才图谱生成——与我们的搜索、匹配产品同一套引擎。想要 OpenAI 的完整名单与可联系的候选人，或者为你正在竞争的某家公司生成同款透视？留个联系方式，我们 1 个工作日内对接。This report is generated from the Metix AI talent graph — the same engine behind our search and matching products. Want the full OpenAI list with contactable candidates, or the same X-Ray for a company you're competing against? Leave your contact details and we'll be in touch within 1 business day.