Bloomberg Fulltime SDE Interview Guide and Question Bank

Hello friends, First, I want to reveal a recruitment truth to my Chinese friends: in the US, over 50% of interviewees cheat. Some use AI, but most have friends or groups who have interviewed them befo

Hello friends, First, I want to reveal a recruitment truth to my Chinese friends: in the US, over 50% of interviewees cheat. Some use AI, but most have friends or groups who have interviewed them before and given them questions. Traders at Berkeley has a question bank of all quant companies and regularly shares old interview questions. The Georgia Tech fraternity has old question banks of all major tech companies, UIUC has its own club, and so on. So for companies like Palantir, Databricks, Stripe, and Bloomberg, if you have a friend who has interviewed you before, it's almost guaranteed you'll get an offer. To get the same opportunity, I want to be your friend. Bloomberg's interview process typically consists of five stages. The first round is an hour-long technical interview, including self-introduction, some basic behavioral questions, and a LeetCode-style problem, requiring optimization afterwards. The second round is also an hour-long technical interview, with two medium-difficulty LeetCode problems, starting with a brief discussion of the resume, followed by the technical part. The third round is also a technical interview, with two questions, but the format may lean more towards data structures or implementation problems. The fourth round is an engineering manager interview, in the form of an in-depth project discussion. Candidates need to explain in detail the projects they've worked on and the underlying thought processes. This round focuses more on technical background and project understanding than on-the-spot coding. The fifth and final round is an HR interview, mainly a relaxed conversation about motivation, interests, and the candidate's expectations for the position. The entire process typically includes self-introduction, resume discussion, technical challenges, and a final Q&A session. After each round of interviews, HR usually provides feedback and arranges the next steps within one to two weeks. Question Bank: I know many people don't have Leetcode Premium, so I've compiled a question bank containing all the questions we've received in Bloomberg interviews so far. My friends have all succeeded using it; all the questions they received are there. This year it's updated; all the questions from the second and third rounds are included. https://leetcode.com/problem-list/an6ko651/ This is an internal question bank that Bloomberg requires employees to use in interviews. This is NDA information, so please do not share it outside of this platform. — Title: Trading Symbols Set with O(1) Ops Difficulty: Medium Prompt: Design a data structure for a trading platform to manage a dynamic set of stock symbols. Support: add(symbol), remove(symbol), and getRandom() that returns a uniformly random symbol from the current set—all in average O(1) time. Do not use built-in random-set libraries. ⸻ — Title: Design Commuter Cost Tracker for Branches Difficulty: Medium Prompt: Build a system to track average travel times (costs) between company sites (e.g., “JFK→Manhattan”, “Boston→NYC”) for employees traveling on expense accounts. Implement checkIn(id, siteName, time), checkOut(id, siteName, time), and getAverageTime(startSite, endSite) to return the average duration between two sites across all trips completed. ⸻ — Title: Decode Compressed FIX Message Difficulty: Medium Prompt: Given an encoded trade message string where patterns like “3[AB]2[C]” represent repeated fields (e.g., for FIX-style compression), decode it into its expanded form. Nested encodings are possible and must be fully expanded. ⸻ — Title: Minimum Rooms for Client Meetings Difficulty: Medium Prompt: Given a list of start/end times for client meetings scheduled at a bank office,

return the minimum number of meeting rooms required so no meetings overlap. Times are half-open intervals. ⸻ — Title: Count Trading Venue Clusters Difficulty: Medium Prompt: You’re given a grid representing branch connectivity where ‘1’ = connected trading terminals and ‘0’ = offline. Count how many isolated clusters (connected components via up/down/left/right) of online terminals exist. ⸻ — Title: Assign Analysts to Two Cities (Cost Minimization) Difficulty: Medium Prompt: Each analyst must be sent to either City A or City B for a rotation. You’re given costs for each analyst to go to A and to B, and exactly half must go to each city. Minimize total travel cost and return that minimum. ⸻ — Title: Clean Canceling Trade Pairs in Log II Difficulty: Medium Prompt: Given a trade action string, repeatedly remove groups of k adjacent identical characters (e.g., k identical “B” = buy corrections) until no more removals are possible.

Return the final string (the cleaned trade log). ⸻ — Title: Merge Maintenance Windows Difficulty: Medium Prompt: Given a list of market data maintenance windows (time intervals) on trading servers, merge all overlapping intervals and return the consolidated schedule. ⸻ — Title: Longest Unique Client Session Difficulty: Medium Prompt: Given a session log string where each character represents a client token,

return the length of the longest substring without repeating tokens, representing the longest continuous window of unique client activity. ⸻ — Title: Design Market Data Ordered Stream Difficulty: Easy Prompt: Implement an ordered stream for market data chunks arriving with ids from 1..n. The stream supports insert(id, value) and returns the longest consecutive sequence starting at the smallest yet-to-emit id after each insert. ⸻ — Title: Flatten Nested Fund Structures (Doubly Linked) Difficulty: Medium Prompt: You are given a multilevel doubly linked list representing nested fund-of-funds relationships. Flatten the list so that all nodes appear in a single-level doubly linked list, preserving depth-first order. ⸻ — Title: Flag Invalid Card Transactions Difficulty: Medium Prompt: Given a list of credit card transactions with {name, time (minutes), amount, city},

return all transaction strings that are invalid. A transaction is invalid if amount > 1000 or if the same name has transactions in different cities within ±60 minutes. ⸻ — Title: LRU Cache for Quotes Difficulty: Medium Prompt: Implement an LRU cache for real-time quotes supporting get(key) and put(key, value) with capacity limit. If inserting exceeds capacity, evict the least recently used quote. Both operations should be O(1) average. ⸻ — Title: Keyword Search in Compliance Grid Difficulty: Medium Prompt: Given a 2D grid of letters representing a scanned compliance document and a target keyword, determine if the keyword exists in the grid by sequentially adjacent letters (up/down/left/right). Cells cannot be reused in a single match. ⸻ — Title: Verify Ticker Code Anagrams Difficulty: Easy Prompt: Given two strings representing ticker codes or portfolio tags, determine if one is an anagram of the other by comparing character counts. Ignore letter case if specified. ⸻ — Title: Validate Price Tree Structure Difficulty: Medium Prompt: Given a binary tree representing a hierarchical price range book, verify it is a valid binary search tree (BST) where all left subtree keys < node key < all right subtree keys for every node. ⸻ — Title: Segment Concatenated Financial Terms (All Sentences) Difficulty: Hard Prompt: Given a string s and a dictionary of financial terms,

return all possible sentences where s The output is segmented into a space-separated sequence of valid terms. The order of output does not matter. — Title: Design P&L Leaderboard Difficulty: Medium Prompt: Implement a leaderboard to track trader scores (e.g., daily P&L). Support addScore(id, scoreDelta), top(K) to sum the top K scores, and reset(id) to set a trader's score to 0. Ensure operations are efficient for large user counts. That's it. I believe that with these resources, as long as you are well-prepared and strive to express your ideas and answer behavioral questions during the interview, you will get an offer. Have a great interview and good luck!