2025–2026 Bloomberg New Grad Interviews: 98 Experiences (Part 4 :)

*Continued from Part 3

Coverage: Timelines 37–50 of 50 detailed candidate reports, plus 48 shorter claim-specific reports used throughout the full series. The 98-experience count is deduplicated across posts, comments, follow-ups, and cross-posts.

50 detailed candidate timelines, loops, and screens — continued

37: Collaborative coding, then a surprisingly technical manager round

Report date: 2025 (exact month not preserved) · Role: Entry-level SWE · Location: United States · Result: Rejected

The virtual loop contained two collaborative coding rounds, a recruiter conversation, and a manager final. The candidate expected the last round to feel like another fit chat. Instead, it added system design, language tradeoffs, object-oriented follow-ups, and deeper questions about decisions in past work.

The candidate still reached the end of a one-month process and was rejected. Their advice was to prepare medium and harder coding, know every resume project and the reasons behind the stack, and review basic design and OOP even for an entry-level manager conversation.

38: One hard problem, two full implementations, and a two-day rejection

Report date: April 2025 · Role: New Grad SWE · Result: Rejected

The task was Longest Substring with At Least K Repeating Characters. The candidate first proposed divide and conquer, then recognized sliding window but considered it harder to implement. They wrote divide and conquer while explaining extensively and dry-ran two extra test cases.

With about 15 minutes left, the interviewer asked for the sliding-window implementation too. The candidate completed it and was rejected two days later. The report supports the need to understand multiple approaches, but not the author’s speculation that Bloomberg always expects two optimal solutions in forty minutes.

39: Missing Ranges plus every maximum-length substring

Report date: April 2025 · Role: New Grad SWE · Status: Not disclosed

After a short background conversation, the candidate solved Missing Ranges and Longest Substring Without Repeating Characters. The second follow-up changed the output from one maximum length to every substring tied for that maximum.

The candidate stored left/right endpoint pairs in a set, then rebuilt the substrings in a final pass. The author also reported that screen recording was required. That procedural detail did not appear consistently elsewhere, so it should be confirmed for the specific interview rather than assumed.

40: Four coding problems per onsite, final manager, no offer

Report date: March 2025 · Role: 2025 New Grad SWE · Result: Rejected

The candidate scheduled in early January, completed a two-problem phone screen on January 22, and had a three-round onsite on February 12. Each of the first two onsite rounds used ten minutes for resume and 45 minutes for two coding problems. The candidate reported completing all four. HR was behavioral.

The February 25 manager round explored internship details and behavior for about 40 minutes. The candidate emphasized genuine interest in finance, economics coursework, Bloomberg’s communication culture, preferred teams, and AI. They believed there was no significant technical miss. Follow-up emails went unanswered until the March 10 rejection. Community theories about headcount were not verified.

41: Manager round without system design, offer one week later

Report date: March 2025 · Role: New Grad SWE · Result: Offer

The manager began by answering two candidate questions, then asked about favorite programming language, what the candidate wanted from the next job, fears about joining, important team qualities, and personal interests. Each topic expanded into a long “why?” conversation.

There was no system design and little resume drilling, although the author warned that other Bloomberg managers did ask it. The candidate had an undergraduate IT background and two SWE internships, most recently at a healthcare company. They had disclosed another offer deadline during HR. Bloomberg emailed first to arrange an update call; it was not an unexpected cold call.

42: London knockout loop with system design in both later stages

Report date: February–March 2025 · Role: New Grad SWE · Location: London · Result: Offer

The candidate completed a technical phone screen, then a knockout-style virtual onsite with one technical, one system-design, and one HR interview. The design prompt involved accepting a client equation, fetching the required data, computing the result, returning it, and then scaling the service massively.

The final engineering-manager round mixed CV projects, algorithms, behavioral discussion, message queues, CDN/caching, and monolith-versus-microservices tradeoffs. The offer arrived two weeks later. This is one of several reports showing that “manager round” can mean a friendly project chat for one candidate and real design depth for another.

43: NYC offer from a non-CS background

Report date: February 2025 · Role: New Grad SWE · Location: New York City · Result: Offer

The first round derived from Design Hit Counter and kept adding O(1) time-and-space follow-ups. Round two used a Find Peak Element concept plus Combination Sum. Round three used a min-stack problem and a Wordle checker. The manager asked the candidate to choose a project, explain technical decisions, teach unfamiliar ML concepts, and defend evaluation metrics. HR covered motivation and role expectations. The verbal offer came after a longer wait and follow-up email.

44: Word Search, Collatz, browser history, then an unexplained rejection

Report date: February 2025 · Role: 2025 New Grad SWE · Result: Rejected

VO1 combined Word Search with run-length encoding. VO2 used the Collatz process and nested-string decoding. VO3 designed browser history and added a follow-up the candidate felt they handled poorly. HR asked why Bloomberg, how the candidate prioritized jobs and offers, and for a proud project.

The loop began in mid-January with about a week between rounds; VO3 and HR happened together. Technical results arrived within three business days. Project discussion consistently used a proud-project deep dive. The candidate could not obtain feedback identifying whether the browser-history follow-up, HR, or something else caused the rejection.

45: Two technical VOs and HR in one afternoon

Report date: February 2025 · Role: New Grad SWE · Status: Advanced to manager scheduling

The 11 a.m. round began with a challenging project and a harder parentheses-depth variation. The candidate’s first approach was wrong, but interviewer guidance led to a correct solution. A simple BFS task followed.

At 1 p.m., a more serious interviewer asked for a difficult technical example from the candidate’s experience, then used a Bloomberg-specific LRU variation with many edge cases. A Meeting Rooms question was added near the end; only the approach and one test case were required. HR began at 2:30 p.m. and stayed conversational. The candidate moved on to manager scheduling.

46: Candy Crush changed late, then a terabyte palindrome stream

Report date: February 2025 · Role: New Grad SWE · Result: Rejected

The candidate quickly implemented a one-dimensional Candy Crush rule after confirming it and dry-running an example. Near the end, the interviewer said the rule should remove an entire contiguous run of at least three, which the candidate believed contradicted the earlier clarification. They rewrote it in five minutes without time to verify.

The harder prompt defined track(char) and isPalindrome() over a potentially terabyte-scale stream, with palindrome queries below O(n). Pointer and DP ideas did not reach a complete answer in 20 minutes. The candidate later considered rolling hash but questioned collision risk. The source contains no verified intended solution.

47: Number of Islands and vertical-order traversal, then next-day passage

Report date: February 2025 · Role: New Grad SWE · Result: Passed phone screen

The hour opened with introductions, challenge/“Why us?” behaviorals, and written prompts copied into HackerRank. Number of Islands came first. An optimization question required one hint, after which the candidate coded and walked through it.

The second problem was a Binary Tree Vertical Order Traversal variation. Another small hint unlocked the approach; the candidate analyzed time and space, coded with 15 minutes left, and dry-ran a case. Execution was not required. The next-step email arrived the following day.

48: London offer across parsing, BFS/DP, large-data design, behavioral, and manager

Report date: February 2025 · Role: Graduate Software Engineer · Location: London · Result: Offer

The first technical focused on data parsing and filtering. The second mixed BFS and dynamic programming. A dedicated design round addressed handling large amounts of data, followed by behavioral interviewing and an engineering-manager round described as “a bit of everything.”

The only preserved behavioral prompt was “Why Bloomberg?” The source does not reveal the detailed inputs or architecture, so its strongest value is corroborating the London five-stage shape and the presence of a dedicated design round before HR/manager.

49: Three coding stages, a long silence, then HR/manager and rejection

Report date: December 2024 · Role: 2025 New Grad SWE · Result: Rejected

The October 18 phone screen used a tree-traversal problem and produced a same-afternoon pass. An October 24 campus round was handwritten; the candidate felt it went poorly but was invited to continue that afternoon. A conflict moved the next round online, where an easy matrix problem was solved on October 30.

Nothing happened for weeks. On December 12, HR scheduled HR and manager conversations for December 19. The author later reported a quick rejection but did not preserve those final questions. They received only one coding task per hour because they talked through requirements and wrote tests carefully.

50: Conference pipeline began onsite and still ended in a later fail

Report date: October 2024 · Role: 2025 New Grad SWE · Result: Passed this round; later structured status Fail

A recruiter reached out after a conference. The first onsite interviewer used ten minutes for resume, then asked for a tree-path sum. Follow-ups required returning the actual path and optimizing it. The second prompt was a complete New York subway-card swipe class beginning with initialization.

The interviewer explicitly said the candidate had passed and HR would provide next steps. The source does not describe the later round that produced the final Fail status. It is a clean example of why one interviewer’s positive close is not the same thing as an offer.

48 shorter candidate reports ~ newest first

These reports have less detail than timelines 01–50.

Short 01: Two hard graph/backtracking questions after a passed first round

Report date: August 11, 2026 · Role: New Grad SWE · Result: Rejected

The candidate passed round one with two medium problems. Round two contained two hard graph/backtracking questions; they reported solving both within 40 minutes but receiving a rejection the next day. The exact prompts were not disclosed.

Short 02: Tricky sorting problem completed after hints

Report date: August 11, 2026 · *Role: New Grad SWE · Result: Rejected

This independent candidate passed round one, then received a sorting-based medium whose key trick was difficult to discover under pressure. They reached a solution after several hints in about 30 minutes and were rejected the following week.

Short 03: Van Eck sequence in a dismissive virtual onsite

Report date: April 3, 2026 · Role: New Grad SWE · Result: Rejected

A month-long process ended with a virtual round on the Van Eck sequence. The candidate became tangled in index handling, finished the code, then spent about 15 minutes defending it to an interviewer they found dismissive. The recruiter called roughly a week and a half later with the rejection.

Short 04: Rounds 2, 3, and HR on one day

Report date: February 28, 2026 · Role: New Grad SWE · Location: New York City · Status: Awaiting next steps

The candidate completed rounds two and three plus HR on the same day, with roughly 30–60 minute gaps. HR said the recruiter would follow up rather than announcing immediate engineering-manager advancement; the source did not preserve the later outcome.

Short 05: The same superday stages split across two weeks

Report date: February 28, 2026 · *Role: New Grad SWE · Status: Advanced to engineering manager

Another candidate reported the opposite schedule: round two happened first, round three followed one week later, HR occurred immediately afterward, and the engineering-manager interview was scheduled for the next day. The final result was not disclosed.

Short 06: Tokyo next-word prediction round

Report date: February 9, 2026 · Role: 2026 New Grad SWE · Location: Tokyo process · Status: Waiting

Two Hong Kong-based interviewers skipped resume discussion and asked for a most-likely next-word predictor trained from token sequences. Follow-ups challenged storage, alternative data structures, vocabulary size, and extending one-word prediction into an entire sentence.

Short 07: Four remembered patterns from a successful process

Report date: February 7, 2026 · *Role: New Grad SWE · Result: Passed

The candidate’s retrospective preserved four tasks: Longest Substring Without Repeating Characters, Flatten a Multilevel Doubly Linked List, Insert Delete GetRandom, and Decode String. They stressed edge cases and recognizing the underlying pattern quickly.

Short 08: Two campus days followed by recruiter silence

Report date: February 1, 2026 · *Role: New Grad SWE · Stage: Campus onsite · Result: Recruiter stopped responding

Day one used one quickly solved problem plus a second stack-friendly task whose identity was not recoverable. Day two used a grid with obstacles and fuel cells, making fuel-aware reachability the real challenge. A third online round was promised, but the author later said the recruiter ghosted them.

Short 09: Strong first round, HR, then rejection before the manager round

Report date: January 22, 2026 · *Role: 2026 New Grad SWE · Location: New York City · Result: Rejected

Round one mixed 10–15 minutes of internship discussion with one hard tagged problem and a verbal follow-up; the candidate advanced the next day despitespite initially misstating complexity. Later rounds moved from an easy/medium task to a hard task with increasingly demanding optimization follow-ups, then HR. The rejection arrived before an engineering-manager round.

Short 10: Tokyo screen finished with half an hour left

Report date: January 21, 2026 · Role: 2026 New Grad SWE · Location: Tokyo · Status: Waiting

After roughly ten minutes on the resume, the candidate solved Meeting Rooms II and a second task whose obfuscated identifier could not be resolved. They finished both with about 30 minutes remaining; no later result appeared in the source.

Short 11: A 100-minute Texas Hold’em design conversation

Report date: January 16, 2026 · *Role: New Grad SDE · Status: Outcome not disclosed

A scheduled one-hour interview with a team lead lasted about 100 minutes. It covered introductions, internship experience, and object-oriented design for Texas Hold’em poker. The interviewer added several corner cases, but the source did not preserve the required classes, methods, or rules.

Short 12: Engineering-manager debugging instead of system design

Report date: December 12, 2025 · *Role: New Grad SWE · Stage: Final engineering manager · Result: Rejected

The manager verbally described a single-machine file service, gradually introduced failures, and asked how the candidate would debug each situation. The candidate had been told to prepare system design, but described this as a general debugging exercise; a later recruiting discussion established that the completed loop did not produce an offer.

Short 13: Current-process overview ending in an offer

Report date: December 3, 2025 · Role: New Grad SWE · Result: Offer

The candidate described an on-campus first round, then a virtual onsite containing two technical interviews and a 30-minute HR round with breaks between them. Several weeks later, the final manager interview became a deep project discussion. Bloomberg scheduled the offer call two days afterward.

Short 14: Word Search and Trapping Rain Water

Report date: November 23, 2025 · Role: New Grad SWE · Stage: Phone screen · Result: Passed

The candidate solved Word Search with DFS and Trapping Rain Water with two pointers. They specifically warned that a BFS attempt for Word Search could consume excessive memory.

Short 15: In-person rate-limiter round that ran out of time

Report date: November 19, 2025 · *Role: New Grad SWE · Stage: First technical · Result: Rejected

Two campus interviewers opened with resume questions and “Why Bloomberg?” The candidate completed an easy first problem, then received a rate-limiter task with an added condition. They could not finish the implementation despite hints before time expired.

Short 16: Thirty-minute recruiter behavioral scree

Report date: November 18, 2025 · *Role: 2026 New Grad SWE · Stage: Recruiter/behavioral · Result: Passed

The Google Meet covered graduation timing, a two-minute background walkthrough, why Bloomberg, teamwork, and an example of exceeding expectations. The candidate used a scope-control story and described Bloomberg as a large real-time data system rather than only a finance company.

Continued in Part 5 with Short 17–48.