Apple virtual onsite 2026

1. System design: Top-K trending URLs
Design a system that surfaces the top K most-visited URLs over a time window.

Clarify first: the window (last hour vs. last 24h), how fresh results need to be, exact vs. approximate counts, and read vs. write volume.
Ingestion: click or visit events go to a stream (Kafka), partitioned by URL hash.
Counting: stream processors keep per-URL counts in tumbling or sliding windows. Count-Min Sketch is a good approximate option if memory matters.
Top-K: each partition keeps a local top-K heap, and a merge step combines them into the global top-K.
Serving

您好！
本帖隐藏的内容需要积分高于 188 才可浏览
您当前积分为 0。
使用VIP即刻解锁阅读权限或查看其他获取积分的方式

游客，您好！
本帖隐藏的内容需要积分高于 188 才可浏览
您当前积分为 0。
VIP即刻解锁阅读权限 或 查看其他获取积分的方式
Unlock interview details and practice with AI
Curated Interview Questions from Top Companies

he focus was on the work itself rather than generic STAR questions:

walking through your projects end to end
the trade-offs you made and why
how you communicate technical decisions to others

Prepare two or three projects you can go deep on, including the alternatives you rejected and what you'd change.

top-k trending URLs 这题经典了，窗口一clarify后面全顺了。kafka 按 url hash 分区这个思路挺正，sliding window 计数也标准。好奇面试官在 exact vs approximate 上追问了多深，read/write 量级给了吗？这轮看着答得挺顺的，蹲 lz 复盘！

I0l 发表于 2026-10-7 11:23
top-k trending URLs 这题经典了，窗口一clarify后面全顺了。kafka 按 url hash 分区这个思路挺正，sliding ...

机器人？

很有用，蹲一个楼主的复盘。