量化交易公司 Jane Street 宣布将对软件工程师招聘面试进行小幅调整,旨在应对大语言模型(LLM)对软件开发方式的影响 1。鉴于过去几年间观察到 LLM 已改变软件构建方式,且工程师日常广泛使用 LLM 进行编码、调试和审查,该公司计划从今年夏天开始,在校招候选人的现场面试中试行新的评估模式 1。
新试点的面试环节减少了代码编写比重,转而更侧重高层级的软件设计、验证代码质量与结构,以及使用 LLM 产出代码的能力 1。尽管部分环节引入 AI 工具考察,但大多数面试轮次仍保持无 LLM 辅助的传统编码面试 1。Jane Street 明确表示,其核心目标并非寻找具备高度成熟 AI Agent 工作方法论的候选人,而是发掘那些拥有构建优秀软件本能的人才 1。此外,针对经验丰富的求职者,面试风格将更加多样化,涵盖系统设计及过往项目讨论 1。
Jane Street, the quantitative trading firm, is piloting minor adjustments to its software engineering interview process to address how large language models (LLMs) have changed software development practices 1. The company observed over recent years that engineers routinely use LLMs for coding, debugging, and code reviews, prompting a shift in what skills are assessed during recruitment 1. Starting this summer, the firm began testing new formats within onsite interviews specifically targeting campus hiring candidates 1.
The revised interview structure reduces the emphasis on writing code from scratch while increasing focus on high-level software design, verifying code quality and structure, and utilizing LLMs to generate code 1. Despite these changes, the majority of interview rounds will continue to feature traditional coding challenges without LLM assistance 1. Jane Street clarified that it does not require candidates to possess highly mature AI agent workflows; instead, the core objective remains identifying individuals with strong instincts for building good software 1. For experienced candidates, interview styles remain diverse, encompassing system design discussions and reviews of past projects alongside any new elements 1.
评论
还没有评论,欢迎留下第一条。