RepostedDocker logo

Docker · Phone screen

In process

Docker SWE Manager and Coding Rounds: 1 Experience

Software engineer · Full-time4 years experienceDoctorateReported Nov 16, 2026
Shared byjanebeingjane

A 2026 Docker developer-tools engineer's first stage: a dense hiring manager round and an AI-allowed coding round on a per-client rate limiter.

At a glance

  • Process: hiring manager round (1 hour) → coding round with AI allowed (1 hour); design and final rounds still ahead (report 01)
  • Asked recently: a product-manager disagreement question, a per-client rate limiter with a million-client follow-up (report 01)
  • Difficulty: the candidate rated it average and found the coding round comfortable
  • Practise first: Per-Client API Rate Limiter

Coverage: One software engineer with about four years of experience and a PhD, changing jobs, interviewing for a senior role on a developer tools team. The report covers the first stage: a hiring manager round and a coding round. The rounds were on November 13 and 16, 2026, and the post went up on November 16, 2026.

What came up

Stage one had two one-hour rounds. The manager round was dense behavioral and experience questions. The coding round was a per-client rate limiter on the candidate's own editor, with AI tools explicitly allowed. A design round and a final round were still ahead when the candidate posted.

Topic Reports Last seen
Hiring manager round: behavioral questions and experience deep dives 1 of 1 Nov 2026
Coding with AI allowed: a per-client API rate limiter, then a version for a million clients 1 of 1 Nov 2026

Questions we found

Algorithms / Data Structures

Question Reports Last seen
Implement a rate limiter that caps requests per client ID; follow-up: cut the memory needed for a million clients (practice) 1 Nov 2026

Behavioral (BQ)

Question Reports Last seen
A product manager asks for a requirement you strongly disagree with: what do you do, and do you have an example? (practice) 1 Nov 2026

1 detailed candidate timeline, loops, and screens

01: Experienced SWE (about four years), stage one

Interview dates: November 13 and 16, 2026 · Report date: November 16, 2026 · Role: Software engineer · Degree listed: PhD · Years of experience: About 4 · Candidate's rating: Positive, average difficulty · Result: Waiting to hear back

Hiring manager round (November 13, 1 hour)

The hiring manager leads the team. The candidate learned that the team is fully remote with optional offices, has two engineers with three more joining and a goal of eight, and holds a standup every morning, which the candidate felt leaned toward close supervision. The questions were mostly behavioral and deep dives into past work, asked at a fast pace. The one that stumped them: if a product manager asks for a requirement you clearly disagree with, how do you handle it, and do you have a story? Their current team has no product manager, so they said so and offered a related story, but the interviewer skipped ahead. Their advice to others is to prepare a disagreement story with a product manager or stakeholder even if you don't work with one now. Practice it as a product manager asks for a requirement you disagree with.

Coding round (November 16, 1 hour, two interviewers from other teams)

There was no shared editor: the candidate shared their screen and coded in their own editor, and the interviewers said AI tools could be used for tests or code. The task was a rate limiter for an API that enforces a limit for each client ID. The candidate first compared the usual options (token bucket, fixed windows, and sliding windows kept either as timestamp logs or as counters) and built the timestamp-log version, which is exact but stores every timestamp. That took about 30 minutes, and AI-generated tests covering bad input, edge cases and the limit logic all passed. The follow-up asked how to handle storage with a million clients, which led to the counter-based sliding window. They got stuck partway through the estimate comparison, used AI to finish that part, and completed it. Only two minutes were left for their own questions. FastPrep's practice version is Per-Client API Rate Limiter.

Result

The candidate found the round comfortable and says the interviewers cared more about reasoning and trade-offs than about typing every line. Results were pending when they posted.

Keep practicing

Sources: Anywhere useful interview info can be found.

0 total reactions. Your reactions: none.
0 comments

Comments

0

No comments yet — be the first to share what you know.