谁在抢人，谁在失血Who's winning talent, who's bleeding it
30 家金融科技公司的工程人才动量榜An engineering-talent momentum leaderboard of 30 fintech companies

01核心结论Key takeaways

统计对象 = 30 家头部金融科技公司当前在职、岗位职能为“工程与技术”的员工（已剔除实习与拟入职），全球口径，共 25,860 人。 所有数字均为聚合统计，报告不展示任何个人信息。Scope = current employees at 30 leading fintech companies whose job function is “Engineering & Technology” (excluding interns and incoming hires), global scope, 25,860 people in total. All figures are aggregate statistics; the report shows no personal information.

02招聘动量榜：谁在加速，谁在踩刹车Hiring momentum leaderboard: who's accelerating, who's hitting the brakes

衡量动量最直接的指标，是“当前工程师里有多少是过去 12 个月加入的”。全赛道平均 16.9%（约每 6 人 1 人）。 排序后差距惊人：最快的 Ramp 有三分之一工程师是这一年新加入，最慢的几家近乎冻结。中位在职时长也印证了这一点——Ramp 仅 15 个月，Cash App、Varo 已超过 45 个月。The most direct gauge of momentum is “what share of current engineers joined in the past 12 months.” The sector-wide average is 16.9% (about 1 in 6). Sorted, the spread is striking: at the fastest, Ramp, a third of engineers joined this past year, while the slowest few are nearly frozen. Median tenure tells the same story — just 15 months at Ramp, versus over 45 months at Cash App and Varo.

03净流入 / 失血榜：圈内挖角的赢家与输家Net inflow / bleeding leaderboard: the winners and losers of intra-cohort poaching

把 30 家公司之间的相互流动单独拎出来：一个人现在在 A、过去待过同业 B，就记一条 B→A。 用“从同业挖来的人数 − 被同业挖走的人数”衡量谁在圈内净赢。这是Isolating the flows between the 30 companies: if someone is now at A and previously worked at peer B, we record one B→A. We measure who wins net within the cohort as “people pulled from peers − people lost to peers.” This is a方向性信号directional signal——只统计同业间的换手，不代表招聘总量，绝对数字偏小，重点看排序与方向。 — it counts only intra-cohort moves, not total hiring; the absolute numbers run small, so focus on the ranking and direction.

最粗的几条“人才管道”（现员工里来自该同业的人数）：Stripe 既是最大的“出水口”也是最大的“进水口”， 是整个赛道的人才中转站；而 Robinhood 像一所训练营，工程师源源不断流向 Coinbase、Ramp、Stripe。The thickest talent pipes (count of current staff who came from that peer): Stripe is both the biggest outflow and the biggest inflow — the sector's talent clearinghouse; meanwhile Robinhood acts like a training camp, steadily feeding engineers to Coinbase, Ramp, and Stripe.

04两个金融科技世界：大厂系 vs 区域系Two fintech worlds: Big Tech-bred vs regionally bred

“工程师从哪里来”把这 30 家公司清晰地分成两半。下图是每家公司里“有过美国大厂（FAANG 级）经历”的工程师占比—— 美国公司普遍很高，欧洲 / 拉美公司普遍很低。但低不等于“没背景”，而是他们从另一套人才池招人。“Where engineers come from” splits these 30 companies cleanly in two. The chart below shows each company's share of engineers with US Big Tech (FAANG-tier) experience — uniformly high at US companies, uniformly low at European / LatAm ones. But low doesn't mean “no pedigree”; it means they hire from a different talent pool.

05创始人外溢：谁是“创始人摇篮”，他们去了哪Founder spillover: who is the “founder cradle,” and where they went

把视角拉长到校友：这 30 家公司累计走出 3,829 个创始人 / CEO 席位（在多家任职者分别计入）。 关键问题是“去了哪”——只有Zooming out to alumni: these 30 companies have produced 3,829 founder / CEO seats in total (counted separately for those who served at multiple firms). The key question is “where did they go” — only 19% 又创办了金融科技公司，其余八成把经验带去了更广阔的创业经济。论人均密度，crypto 与第一代独角兽的校友最爱创业。went on to found another fintech, while the remaining four-fifths took their experience into the broader startup economy. On per-capita density, alumni of crypto and first-generation unicorns are the keenest founders.

0630 家全景对照表The 30-company side-by-side table

一张表看完所有维度。“工程占比”= 工程师占公司总人数之比（体量不等于工程实力：Kraken 高达 60.6%，Revolut 仅 11.0%）； “近一年入职”= 招聘动量；“圈内净流动”= 同业挖角净值；“大厂背景”= 有 FAANG 级经历占比；“造老板/千人”= 创始人外溢密度。Every dimension in one table. “Engineer share” = engineers as a share of total headcount (size ≠ engineering depth: Kraken hits 60.6%, Revolut just 11.0%); “joined in the past year” = hiring momentum; “net intra-cohort flow” = net peer poaching; “Big Tech background” = share with FAANG-tier experience; “founders per thousand” = founder-spillover density.