Changelog

Microsoft joins the catalog: 39 questions reconstructed from 93 candidate reports spanning October 2025 to October 2026, across Azure, Microsoft AI, Copilot, hiring events and other teams: online assessment problems, phone screens, onsite coding rounds, ML coding such as beam search, system design and the behavioral rounds. A Microsoft AI tab collects the questions Microsoft AI teams ask. Thirty-one questions ship with a solution executed against its own test cases (248 cases, all passing). Two are free to read: the DNA fragment assembly screen and Merge Intervals. Included with any subscription, or on its own with a one-month Microsoft pass.

DoorDash joins the catalog: 12 questions reconstructed from 41 candidate reports spanning April to October 2026: the Code Craft phone screens (Dasher Pay and the Bootstrap API), the AI Code Craft and Debugging onsite rounds, food review and donation event system design, the project deep dive, incident scenarios and the hiring manager round. Seven questions ship with a solution executed against its own test cases (77 cases, all passing), and both design rounds have a full walkthrough. Two are free to read: the Bootstrap API phone screen and the food review design. Included with any subscription, or on its own with a one-month DoorDash pass.

Uber joins the catalog: 83 questions reconstructed from 163 candidate reports spanning October 2025 to October 2026: 25 phone screen and 26 onsite coding questions, 14 online assessment problems, 10 Uber and Uber Eats system design rounds, 6 ML Engineer rounds, the project deep dive and the behavioral round. Seventy questions ship with a solution executed against its own test cases (569 cases, all passing), and every system design round has a full walkthrough. Three are free to read: an online assessment problem, the robot-distance phone screen and the parking lot design onsite round. Included with any subscription, or on its own with a one-month Uber pass.

Robinhood and Roblox add 25 candidate reports posted from August to October. On Robinhood, the Job Scheduler design rises to 26 reports with four new interviewer follow-ups, and Fractional Share, Referral Count and the project deep dive each gain four or five. On Roblox, the online assessment games, Candy Crush and the most-frequent call path gain new reports. Five Roblox online assessment problems are new: Lasers and Robot, Newspaper Text Alignment, Digit-Swap Pairs, Library Book Circulation Tracker and Stepwise Structures, four of them written up from candidates’ screenshots. Both catalogs now open on the solve-first workspace: 15 Roblox and 9 Robinhood coding questions, each checked against its source reports and graded by its own tests.

Snowflake moves to the solve-first workspace. Every coding question was checked again against its source reports: 86 questionsfrom 125 candidate reports now, after thin pages and questions from other companies’ interviews were removed and duplicates merged. Sixty-nineship with a solution executed against its own test cases (503 cases, all passing), including five online assessment problems written up from candidates’ screenshots. Each question’s reported outcome now comes from what candidates said about that round. The free previews are Closest Cake, Basic Calculator II, the Dynamic Programming OA and the Distributed Job Scheduler design.

Capital One joins the catalog: 38 questions reconstructed from 48 candidate reports spanning October 2025 to October 2026: 32 online assessment problems, the Power Day banking coding round, the virtual card case study, credit card system design, the behavioral round, and the MLE and Applied Researcher technical rounds. Thirty-fourquestions ship with a solution executed against its own test cases (194 cases, all passing), including the case study’s debugging part. Four are free to read: two online assessment problems, the behavioral round and the MLE technical rounds. Included with any subscription, or on its own with a one-month Capital One pass.

Ramp joins the catalog: 14 questions reconstructed from 28 candidate reports spanning November 2025 to October 2026: the URL Maze tech screen, 3 onsite coding, the AI-assisted coding round, the presentation and system design rounds, behavioral, and 6 online assessment pages including the two frontend assessments. Eight questions ship with a solution executed against its own test cases (64 cases, all passing), and both design rounds have a full walkthrough. Three are free to read: the Task Manager online assessment, the flight history coding round and the frontend hidden-URL assessment. Included with any subscription, or on its own with a one-month Ramp pass.

Amazon joins the catalog: 57 questions reconstructed from 208 candidate reports spanning October 2025 to October 2026: 11 online assessment, 3 phone screen, 20 onsite coding, 8 design, 2 Applied Scientist ML and 13 Leadership Principle pages. The online assessment pages cover the 2026 format, an algorithm problem followed by about an hour of debugging a repository with an AI assistant, and the two reported repositories are rebuilt as reconstructions written by AceOffer, which the pages say. The Leadership Principle pages collect the reported wordings of each question and what interviewers probed. Thirty questions ship with a solution executed against its own test cases (272 cases, all passing), and all eight design rounds have a full walkthrough. Course Schedule, the most reported onsite coding question, is free, and so was every online assessment page until October 17, when two stayed free as samples. Included with any subscription, or on its own with a one-month Amazon pass.

Rippling joins the catalog: 17 questions reconstructed from 118 candidate reports spanning October 2025 to October 2026: 5 phone screen, 6 onsite coding, 4 system design and 2 behavioral. Rippling tests practical, multi-part coding rather than LeetCode, and many coding rounds let you use an AI assistant. Eleven questions ship with a solution executed against its own test cases (74 cases, all passing), and all four system design rounds have a full walkthrough. The music player round is a reconstruction written by AceOffer, and the page says so. The free preview is Delivery Driver Payment, the most reported phone screen. Included with any subscription, or on its own with a one-month Rippling pass.

Cursor joins the catalog, and every page is free: 6 questions reconstructed from 14 candidate reports spanning May 2025 to September 2026 — 3 phone screen coding problems and 3 system design rounds. The signature phone screen builds a Merkle tree over a real repository, then diffs two trees, then syncs a client to a server; ten reports describe it. All three coding questions ship as part-by-part problems with a solution executed against its own test cases (22 cases, all passing). The design pages follow the brief and the follow-ups candidates reported; the answers are standard preparation, and the pages say so. The internationalization design comes from two reports of Graphite rounds, the code-review startup Cursor acquired.

Scale AI joins the catalog: 17 questions reconstructed from 49 candidate reports spanning January 2025 to August 2026 — 4 phone screen, 4 onsite coding, 4 ML, 2 system design, 2 behavioral and a project presentation. Scale AI tests practical work rather than LeetCode: a phone screen that has rotated from poker hand rules to party schedules to a task scheduler, a debugging round in an unfamiliar codebase, a backend practical that calls an LLM, and design rounds built around black-box model services. Nine questions ship with a solution executed against its own test cases (117 cases, all passing) in the solve-first workspace, and both design rounds have a full walkthrough. The debugging rounds are reconstructions written by AceOffer, and the pages say so. Every page was checked against its source reports, and where reports disagree or leave a detail out, the page says to ask rather than guessing. The free preview is Party Time Blocks, the most reported phone screen. Included with any subscription, or on its own with a one-month Scale AI pass.

Citadel joins the catalog, covering both Citadel and Citadel Securities: 62 questions reconstructed from 90 candidate reports spanning January 2025 to September 2026 — 25 phone screen, 18 online assessment, 11 onsite coding, 3 system design, 3 technical deep dive and 2 behavioral. The problems lean toward trading-flavoured data structures: merging sorted price feeds (the most reported, 6 reports), order books and matching engines, LRU caches and ring buffers, alongside standard LeetCode problems. Fifty ship with a solution executed against its own test cases (440 cases, all passing), and every one opens in the new solve-first workspace: the problem, hints and the editor side by side, one tab per graded part. Every page was checked against its source reports, and where reports disagree or leave a detail out, the page says to ask rather than guessing. The free previews are Merge K Sorted Feeds, Earliest Free Meeting Slot and Distinct Palindromic Substrings. Included with any subscription.

Snowflake joins the catalog: 103 questions reconstructed from 129 candidate reports spanning January 2025 to September 2026. The loop leans harder on the technical screen than most: it is usually two one-hour coding rounds, and several candidates report a system design round inside the screen rather than only at onsite. The problems are standard algorithms rather than company-specific exercises, and a handful recur — topological sort on a task dependency list (16 reports), the cake-and-people distance problem (12), tree height after deleting nodes (11), and building a key-value store (10) — each with the optimisation follow-up interviewers wait for. From 2026 some loops open with a 20-minute AI-conducted screen before any human round. Alongside: 38 phone screen, 38 coding, 12 system design, 11 OA and 4 behavioral questions. Seventy-two ship with a solution executed against its own test cases (369 cases, all passing), and all twelve system design questions ship with a full walkthrough. The free previews are Closest Cake, Basic Calculator II, the Dynamic Programming OA and the Distributed Job Scheduler design. Included with any subscription.

Databricks joins the catalog: 57 questions reconstructed from 221 scraped threads spanning January 2025 to August 2026 — the largest corpus on the source forum after Roblox, and the one whose reports agree most about the question pool: candidates say it is small and repeats, and the top coding questions are reported in close detail across a year. Fifteen coding questions ship with a solution executed against its own test cases, and all seventeen system-design questions ship with full breakdowns. The rounds grade bug-free implementation code — up to 90 minutes, your own test cases expected — and the stage nobody warns you about comes after: a hiring committee and reference checks, with several reports of strong-hire feedback on every round followed by a rejection. The free preview is the CIDR firewall-rules round (frequency 33) with its follow-up ladder and tested solution. Included with any subscription.

Robinhood joins the catalog: 46 questions reconstructed from 94 scraped threads spanning January 2025 to October 2026 — the densest company on the source forum so far. The anchor is the distributed job scheduler, reported more than 3x anything else and sometimes named on the phone-screen invite itself; all seven system-design questions ship with full breakdowns, and every coding question reported three or more times has a solution executed against its own test cases. Design rounds here grade correctness under failure— at-most-once execution, cancel-versus-fill races, worker crashes — and coding rounds run long: reports estimate 150–200 lines, graded on working code over elegant structure. The free preview is the service load-factor DAG round (frequency 14) with its follow-up ladder and tested solution. Included with any subscription.

Zip (ZipHQ) joins the catalog: 17 questions reconstructed from 33 candidate reports spanning February 2025 to July 2026, twelve of them with solutions executed against their own test cases. The loop is front-loaded: reports show candidates rejected at the phone screen more often than at the onsite, and there are frequently two phone screens — a LeetCode-style round, then a practical round inside an existing repository where you fill in business logic and write its unit tests. That practical round is the most-reported question here. The self-serve path is a 90-minute, four-level gated OA where each level’s tests must fully pass before the next unlocks; one candidate cleared three of four, scored 750/1000, and was still rejected. The free preview is an LRU cache debugging challenge whose two planted bugs cancel each other on the obvious test — fix only one and a passing case starts failing. Included with any subscription.

Roblox joins the catalog: 34 questionsreconstructed from 200+ candidate reports across 135 threads spanning February 2025 to August 2026 — the largest corpus here after the two AI labs, and the first non-AI company. Roblox’s question set is unusually concentrated: a sliding-window rate limiter at 29 reports and a most-frequent call-path problem at 22 account for roughly a third of everything reported. Both arrive as graded parts submitted one at a time, and in both the base problem is not what decides the round — the follow-up is. Two separate reports describe passing every visible test on the rate limiter and failing a hidden one for the same reason. The design round is most often like/unlike at scale or delayed payments, and reports reward naming the Redis key and the read and write queries over staying at box-and-arrow level. Included with any subscription.

Harvey AI joins the catalog: 9 questions reconstructed from 50+ candidate reports spanning March 2025 to August 2026. Harvey repeats itself more than any other company covered here — twelve of the fifty reports describe the question they were given as the known one from the forum. The phone screen draws from a small rotating set, most often the RAG citation / text-matching round, which at 23 reports is among the most-reported questions in the entire catalog; the onsite design round is almost always Google Drive-style storage with ACL or a RAG pipeline. Included with any subscription.

Palantir joins the catalog: 18 questions reconstructed from 28+ candidate reports spanning January 2025 to July 2026 — including the two rounds nothing else in the catalog covers: the Learning round, where the interviewer teaches you an unfamiliar system (a single-threaded package manager for SWE, Palantir’s internal SQL-like query language for FDSE) and asks you to extend it, and the open-ended Decomposition round. Candidate reports state directly that Palantir reuses its questions and splits them by track. Included with any subscription.

Mercor joins the catalog: 3 questions reconstructed from 14+ candidate reports — the distributed job-scheduler system-design round (seen for both Marketplace and Frontier Data MLE roles), the signature 3-part referral-network coding round with a harness-tested solution, and the faulty-comparator top-k algorithm round. Every Mercor question is free — free to read. More Mercor questions ship as new reports come in.

Perplexity joins the catalog: 8 questionsreconstructed from 29+ candidate reports — the OA-style multi-part coding rounds Perplexity is known for (in-memory Unix file system, an “Implement ToDo list for AI” task DAG with cascade failure, the stop-words streaming problem), the byte-tokenizer round, beam search, an expiring-credit tracker and a temporal key-value store with restore, plus a codex-fact-checked Design Mint.com system-design walkthrough, all with harness-tested solutions. The in-memory file system question is free — free to read.

xAI joins the catalog: 16 questions reconstructed from 80+ candidate reports — the staged in-memory database with nested transactions (reported 9 times), the complete 15-minute phone-screen script (reported 10 times), a lazy-refill rate-limiter walkthrough, the RAG pipeline design-plus-evaluation round, GPU node failure detection, a concurrent job scheduler, and harness-tested coding solutions throughout. The phone-screen guide and the in-memory database question are free — free to read.

Waymo joins the catalog: 28 questions reconstructed from 55+ candidate reports — real technical phone screens (BFS/graph, pure-Python data processing), AV-flavored system design (evaluation systems, inference at 100M DAU), six coding solutions, and full system-design walkthroughs. The entire Waymo catalog is free — free to read.