Bloomberg AI Interview, Hiring Level Not Reported: 2 Experiences

Shared by janebeingjane ·
Reports
2 interview reports
Level
Unknown
Stages
Phone screen

In fall 2026, Bloomberg AI-assisted coding reports without a hiring level: a HackerRank editor with a required agent, a Dublin loop, and a Django round where manual edits were not allowed.

At a glance

  • Process: entry 02 had two AI-assisted rounds on different days, a virtual round 1 and a virtual onsite round 2; entry 01's stage wasn't given.
  • Asked recently: a single Python file with a test suite, then a larger Python codebase with about five tickets (entry 02).
  • Difficulty: entry 01 finished every task and was still rejected, unsure how much AI use was expected; entry 02 gave no outcome.

Coverage: 2 candidates whose hiring level isn't stated: a software engineer interviewing for the Dublin office, and one who gave no role · AI-assisted coding rounds · fall 2026 (one report posted November 2026, one undated reply)

Other levels: Bloomberg SWE intern AI-assisted coding rounds, Bloomberg new grad AI-assisted coding rounds and Bloomberg experienced engineer AI coding interviews.

What came up

Entry 02 is one of the few reports that describes the tool itself: a VS Code style editor with a coding agent, inside HackerRank, where using the agent was required.

Topic Reports Last seen
Expectations for how much to use the AI 2 of 2 Nov 2026 (report date)
A single file with tests, then a larger codebase with tickets 1 of 2 —

Questions we found

AI Coding

Question Reports Last seen
Solve a task in a single Python file with a test suite, using the agent 1 —
Work through roughly five tickets in a bigger Python project with the agent 1 —

2 detailed candidate timelines

01: Every task done, still a no

Report date: Nov 9, 2026 · Role: Software Engineer, level not stated · Location: Dublin · Result: Rejected

AI-assisted interview

  1. The candidate sat Bloomberg's AI-assisted interview for the Dublin office.
  2. They couldn't work out how much AI the interviewer wanted them to use.
  3. Every required task got done, and the answer was still no. They didn't describe the tasks or say which stage this was.

02: A HackerRank editor with plan, ask and agent modes

Report date: not given (a reply in a thread that opened Oct 12, 2026) · Role: not stated · Result: Not shared

Round 1 (virtual): one file and its tests

  1. A single Python file with a test suite. No framework, and no spec file.
  2. The environment was a VS Code style editor with a Copilot-like assistant, running inside HackerRank. Using the agent was required, not optional.

Round 2 (virtual onsite, another day): a larger codebase

  1. Round 2 moved to a bigger Python project. There were roughly five tickets, and the agent was meant to do the work; the candidate didn't describe the tickets.

How the AI was used and graded

  1. The agent proposed changes; the candidate reviewed each diff, accepted or rejected it, and explained the decision.
  2. In the candidate's view, the grading looked at understanding the code and the question, steering the agent, and making good use of its plan, ask and agent modes.

One more report, possibly a repeat

A report posted on Nov 18, 2026, with no level given, describes a later round in a full-stack Django app. Beforehand the candidate knew only that the codebase would be big and new to them, not which framework it used.

  1. The task was a new feature that touched the frontend, the backend and the database schema.
  2. They asked Claude for an overview of the architecture and a plan, then started editing code themselves. The interviewer stopped them: changes were to be made through the AI, not by hand.
  3. The interviewer also asked about one of their own projects. They felt the first round had gone well and this one less so; no result was given.

The same kind of Django round appears in the experienced engineer report, and this may be the same candidate, so it isn't counted in this article's totals.

Patterns across the reports

  1. How the AI is used seems to count. Entry 01 never learned how much AI use was wanted and was rejected despite finishing; entry 02 knew the agent was required and believed the graders wanted all three agent modes used well.

Common questions

Which AI tool does Bloomberg use in the AI-assisted round?

Entry 02 describes a VS Code style editor with a Copilot-like agent inside HackerRank, with plan, ask and agent modes; it didn't name the product or model. The Nov 18 Django report names Claude.

Can you edit the code yourself in Bloomberg's AI round?

Not always. In the Nov 18 Django report, the interviewer told the candidate to make changes through the AI after they began editing by hand. Entry 02 also says using the agent was required.

Is finishing the tasks enough to pass Bloomberg's AI interview?

Not in entry 01, which finished every task and was rejected. Entry 02 suggests how the agent is steered also counts.

Keep practicing

AI coding rounds at other companies

More companies with AI rounds on FastPrep:

Across companies: 2026 AI Coding OAs at 15 Companies: 32 Reported Problems and 2025–2026 AI Coding Phone Screens and Onsites at 43 Companies: 76 Reported Problems.

Sources: Anywhere useful interview info can be found.

0 total reactions. Your reactions: none.

Comments

0

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