Skip to content

ElevenLabs Full-Stack Engineer Onsite面试经验 ​

发表于 2026-11-10
更新于 2026-11-10

面试详情 ​

ElevenLabs的onsite是4轮:

  • V1: Live coding
  • V2: Product decomposition
  • HM
  • CTO

V1:Live coding ​

V1的问题套了个AI的壳,就是LLM生成的text是want_to_speak,agent说的是spoken_text,want_to_speak永远是对的,spoken_text有可能有repeated characters, words or word phrases,要求返回最长的prefix of want_to_speak

比如:

text
want_to_speak: I can help you book an air ticket
spoken_text: I can can help you book
expected output: I can help you book

第一部分 ​

第一部分是让你review一个junior dev的代码,他用的LCS,先建一个DP matrix,达到LCS length,然后从右到左重建LCS

他的错误一个是他的DP matrix index out of range,还一个是下面这个情况并不work:

text
want_to_speak: I can help you book an air ticket
spoken_text: I can can
output: I can an
expected output: I can

第二部分 ​

第二部分让你去实现这个function_v2。审题很重要,我也是到快结束才回过味儿来,spoken_text只会出现重复,不会漏词也不会替换词,其实就是两个pointer从左到右,遇到不一样的就只移动spoken_text的pointer就好了

V2:Product decomposition ​

V2就真的很开放,比如footlocker的网站,设计一个voice chat system,非常开放,我经常问面试官你觉得这个要怎么样吗?他会回你觉得应该怎么样,不知道是不是挂点,反馈是we felt there were gaps in AI fluency and product design skills that are critical for this role

引用来源