← Back to Directory

Bloomberg interview notes, page 11 of 14

Large Enterprises

Most of these Bloomberg write-ups come from Software Engineers (130), then Product Managers (2) and Data Analysts (1). The most reported rounds are the technical phone screen (100), the behavioral / leadership round (8), the onsite coding round (8) and the system design round (8).

133 interview notes · updated Oct 2026

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg coding round, one problem but with a twist baked in from the start. Pretty clean interview overall, just needed to think past the standard version of the problem.
  • Given an m x n grid of land and water cells, find all islands (4-connected groups of land). Then, given an array of target sizes, return an array where each entry is the count of islands whose size exactly matches the corresponding target.

“It's Number of Islands but you can't just count them, you need to track the size of each one.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg SWE interview with a classic data structures design problem. Pretty straightforward if you've seen it before, but there are enough moving parts to trip you up under pressure.
  • Design a data structure that supports insert, remove, and getRandom operations, all in average O(1) time. Insert should return false if the value already exists, remove should return false if the value is absent, and getRandom should return each element with equal probability.

“The getRandom part is what makes this interesting.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg coding round with a simulation-style problem that looks straightforward until you actually try to implement it cleanly.
  • Given a champagne tower where you pour a certain number of cups into the top glass and each glass overflows equally to the two glasses below it, how full is the glass at a specific row and position?

“Took me longer than it should have to realize you just simulate the whole thing row by row, tracking overflow.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg Data Engineer interview with a finance-heavy analytical question that felt more like a business case than a technical screen. Just the one question but it had a follow-up that changed the whole framing.
  • When analyzing a financial report from an industry you're unfamiliar with, which single metric would you focus on first, and why? Follow-up: if the company is small and private rather than publicly traded, does your answer change, and how?

“The follow-up is where it gets tricky.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg SWE interview, one of the questions was a streaming data structure problem that didn't need to be fully coded out. Interviewer seemed more interested in the approach than a working solution, which was a bit of a relief honestly.
  • Design a simplified stock price stream that supports updating a stock's price and retrieving the top K stocks by price.

“Went with a heap plus a hashmap for tracking current prices, and used lazy deletion to handle stale entries in the heap.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg SWE interview with a linked list problem that looked manageable on the surface but had enough depth to trip you up if you weren't careful about the traversal order.
  • Given a linked list where each node has both a next pointer and a down pointer (the down pointer leads to a nested sub-list), flatten the entire structure into a single list using only next pointers. The order must follow a depth-first traversal, meaning you fully descend into each down list before continuing along the next pointer.

“My first instinct was to just splice recursively and it worked, but I fumbled explaining the time complexity.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg coding round, pretty standard binary search territory but with a twist at the end that I didn't fully think through in time.
  • Given a sorted array of integers and a target value, return the indices of the first and last occurrence of the target. Return [-1, -1] if not found. Must run in O(log n).
  • Follow-up: if new numbers strictly larger than the current maximum keep getting appended to the array, how do you efficiently maintain the first and last occurrence of the target as the array grows?

“Classic binary search variant.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg SWE interview with a data structures design problem that looked simple on the surface but had enough edge cases to trip you up if you weren't careful. The O(1) constraint is what separates people who've seen this pattern from people who haven't.
  • Design a lottery system that supports inserting a new ticket, drawing a random ticket, and removing a specific ticket, all in O(1) average time.

“I went straight to an array for insert and random draw, which is fine, but then stumbled when they asked about removal.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg coding interview, one question the whole session. Classic problem but the full solution has enough moving parts that it's easy to fumble under pressure.
  • Given a string and a dictionary of words, return all possible sentences you can form by inserting spaces into the string such that every word appears in the dictionary. Words can be reused.

“I knew this problem but blanked on how to handle the memoization cleanly.”

View Post

Bloomberg·Software Engineer·Technical Phone Screen

Jun 2026
Bloomberg coding screen for a software engineer role. One question, pretty self-contained, but the implementation details had more nuance than I expected going in.
  • Given a list of token sequences as training data, build a bigram language model that can predict the most likely next token for any given query token. You need to implement both the training logic and the query interface.

“The core idea clicked fast: count how often each token follows another, then just return the argmax.”

View Post