Data Scientist
virtual
· 难度 4/5
Recruiter Screen,technical phone(统计 + SQL + Python),virtual onsite:围绕一个 abuse detection 问题做 case study,ML deep-dive,behavioral,最后和 hiring manager 面试。Trust 团队处理的是 LinkedIn 级别的数据规模。
由 AI 翻译,查看原文
- Walk me through an analysis that changed a product decision.(带我过一遍一次改变了产品决策的分析。)
- Tell me about a model you deprecated. Why?(说说一个你下线的模型。为什么?)
- Describe partnering with engineering on a real-time ML feature.(描述一次你和 engineering 合作做实时 ML 功能的经历。)
- How do you balance precision and recall for an abuse-detection model?(你会怎么在滥用检测模型的 precision 和 recall 之间做平衡?)
Account Executive
virtual
· 难度 3/5
Recruiter Screen,和 sales manager 的视频面试,模拟客户电话,和跨职能伙伴的 behavioral,最后和区域 director 面试。LinkedIn Sales 有很强的咨询式销售文化,onboarding 也很结构化。
由 AI 翻译,查看原文
- Walk me through your largest enterprise deal.(带我过一遍你做过的最大一笔 enterprise deal。)
- Tell me about a customer that nearly churned. How did you retain them?(说说一个差点流失的客户。你是怎么留住他们的?)
- Describe a situation where the buyer changed mid-deal.(描述一次交易进行到一半,买方中途换人的情况。)
- How do you stay credible when the customer knows your product better than you on day one?(当客户第一天就比你更懂你们的产品时,你怎么保持可信度?)
Machine Learning Engineer
virtual
· 难度 5/5
流程非常严格。先电话面试,然后是ML专项技术面。Virtual onsite一共5轮:ML system design(feed ranking)、coding、ML fundamentals、applied ML(feature engineering)、以及behavioral。他们同时考察ML知识的广度和深度。
由 AI 翻译,查看原文
- Tell me about a ML system you designed from scratch. What were the key decisions?(讲讲你从零设计过的一个 ML 系统。关键决策有哪些?)
- Describe a time you improved model performance significantly. What was your approach?(描述一次你显著提升模型效果的经历。你的方法是什么?)
- How do you think about feature engineering for social network data?(你如何思考社交网络数据的特征工程?)
- Give an example of collaborating with researchers to productionize a new technique.(举个和研究人员合作把新技术落地到生产环境的例子。)