2026 Bloomberg AI-Assisted Coding Interview, SWE Intern Codebase Rounds: 5 Experiences
- Company
- Bloomberg LP
- Reports
- 5 interview reports
- Level
- Internship
- Stages
- Onsite
In these 2026 intern reports, Bloomberg's AI-assisted rounds drop candidates into code they didn't write: a spec to check the code against, an issue to trace, or a stack of tickets in a big system.
Coverage: 4 software engineering intern candidates (London, Warsaw, and two who didn't give a location, one of them interviewing through an on-campus process), interviewed October–November 2026 · AI-assisted coding and codebase rounds, plus the system design round that followed in two loops · A separate supplement adds 1 software engineer whose level isn't stated (roles listed in London and Frankfurt, November 2026)
What came up
Three of the four intern reports describe tasks inside an existing codebase rather than writing code from scratch. Two of them describe a step up in size, from a small file or small codebase in round 1 to a large codebase in round 2.
| Topic | Reports | Most recent report |
|---|---|---|
| Tasks inside an existing, unfamiliar codebase | 3 of 4 | Nov 9, 2026 (report date) |
| Small file or small codebase first, large codebase next | 2 of 4 | Nov 9, 2026 (report date) |
| System design as the round after the codebase work | 2 of 4 | Nov 9, 2026 (report date) |
| Ticket-style tasks in a large system | 1 of 4 | Nov 9, 2026 (report date) |
| Finding where code goes wrong and fixing it | 2 of 4 | Oct 29, 2026 |
| Interviewers watching how the AI assistant is used | 2 of 4 | Oct 29, 2026 |
| Spotting ambiguities in a written spec | 1 of 4 | Oct 29, 2026 |
| Testing the fix and explaining it | 1 of 4 | Oct 6, 2026 |
One report can sit in several rows. Report 01 didn't share its task, so it only counts in the size row.
Questions we found
None of these tasks matches a Bloomberg practice exercise on FastPrep, so the rows are plain text.
AI Coding
| Question | Frequency | All last seen dates |
|---|---|---|
| Resolve four tickets in a large matchmaking-system codebase (queues, workers, many business rules) | 1 report | Nov 9, 2026 (report date) |
| Given a codebase and a spec, find the spec's ambiguities and the code that doesn't match it, first in a small codebase, then in a large one | 1 report | Oct 29, 2026 |
| Investigate an issue in an unfamiliar repository, find the root cause, fix and test it, then explain how the AI helped | 1 report | Oct 6, 2026 |
The system design prompts in reports 02 and 03 weren't shared, so they have no row.
4 detailed candidate timelines
01: Round 1 on one small file, a large codebase still to come
Report date: Nov 9, 2026 · Role: SWE Intern · Result: In process (round 2 not yet taken)
Round 1
- The round was built around a single small file. The candidate found it manageable for that reason.
- They didn't share the task.
Round 2 (scheduled)
- The candidate was expecting round 2 to be an AI-assisted round in a large codebase.
- Their worry was the size jump. They weren't sure how to approach a big codebase or exactly how they were expected to use the AI assistant.
02: Four tickets in a matchmaking system, then system design
Report date: Nov 9, 2026 · Role: SWE Intern · Process: On campus at the candidate's school · Result: Rejected after round 3
Round 1 (on campus)
- No content shared. The interviewer asked for availability for the next round as soon as it ended.
Round 2: codebase tickets
- Four tickets in a very large codebase for a matchmaking system, with queues, workers and a lot of business rules.
- The candidate called it hard but said the interviewer was relaxed.
- They solved one ticket in about 30 minutes and a second in about 10. Two out of four was enough: they passed, and the round 3 invite came about three days later.
The candidate's replies don't describe the AI assistant or how they used it in this round.
Round 3: system design
- The first ten minutes were about the resume, one or two light questions.
- Then a very ambiguous design problem. The candidate didn't share it. The expected flow was clarifying questions, API endpoints, a SQL or NoSQL choice with a schema, then a first high-level design, and time at the end for their own questions.
- They spent longer on clarifications than they should have. The interviewer kept pushing them to start designing, and some early parts were wrong, so the design changed several times. They felt the final design was good, but it took a lot of trial and error.
- The rejection came about six days later. Afterwards they found an easier version of their problem on a practice site, labeled hard.
03: A spec, a small codebase, then a large one
Interview date: Oct 29, 2026 · Role: Software Engineer Intern · Location: London · Applied: Through the university · Process: About two weeks, three rounds · Result: No offer
Round 1: AI-assisted codebase interview (small)
- The candidate got a small codebase and a spec.
- The job was to spot where the spec was ambiguous, then find and debug the parts of the code that didn't follow it.
Round 2: AI-assisted codebase interview (large)
- The same kind of task on a large codebase.
- Across both rounds, the interviewers wanted to see the AI used for three things: clearing up the spec's ambiguities, pointing out where the code and the spec disagreed, and finding the way around the codebase.
Round 3: system design
- No details shared.
04: Root-causing an issue in an unfamiliar repo
Interview date: Oct 6, 2026 · Role: SWE Intern · Location: Warsaw · Result: No offer (the candidate still rated the experience positive)
AI-assisted codebase task
- An issue in a repository the candidate hadn't seen before. They had to investigate it, find the root cause, implement a fix and test it.
- Then they explained their reasoning and approach, and specifically how they used the AI assistant to reach the fix and to check that it worked.
This is the only round the report describes.
Supplement: hiring level not reported
This candidate didn't say whether the role was intern, new grad or experienced (they mentioned earlier jobs at two large tech companies), so the report stays outside the intern counts above.
AI Coding (hiring level not reported)
| Question | Frequency | All last seen dates |
|---|---|---|
| Implement the requested features in a large Django (Python) codebase so the provided tests pass | 1 report | Nov 9, 2026 (report date) |
05: A Django codebase and half an hour to work in it
Report date: Nov 9, 2026 · Role: Software Engineer, level not stated · Location: London / Frankfurt roles · Result: Rejected
Round 2: AI codebase interview
This was the candidate's second round. They didn't describe the first.
- The round ran about an hour, but the first 30 minutes were behavioral (no questions named), which left roughly 30 minutes for the code.
- The codebase was large and written in Django, a Python framework the candidate had never used.
- The task was to build the requested features, with test cases to pass. The agent available in the environment ran models the candidate described as Sonnet 5 and GPT 5.6 class, which in their view could easily finish tasks like these in one shot.
- They used the agent's planning and ask modes, put the requirements into their own words, and went one feature at a time. A few tests passed; there was no time to get through every feature.
- Rejected.
Looking back
- Most of the time went into understanding the codebase and finding where each change belonged.
- Reading the agent's output while talking the interviewer through it felt close to impossible in 30 minutes.
- Nobody told them how much AI use was fine. In hindsight they suspected this interviewer wanted the task done in one shot, and they concluded that each interviewer judges AI use differently.
Patterns across the reports
These patterns cover the four intern reports only.
-
The codebase gets bigger in round 2. Reports 01 and 03 both describe a first round on something small (one file in 01, a small codebase in 03) and a second round in a large codebase. Report 03 went through both; report 01 had only been told what round 2 would be.
-
Codebase rounds, then system design. In reports 02 and 03 the codebase work led into a system design round as round 3, and both candidates were turned down after it. Report 02 is the one with detail: a vague prompt, resume questions first, and an interviewer pushing to move from clarifying to designing.
-
Debugging against what the code should do, with the AI use itself on show. Reports 03 and 04 are both about finding where code goes wrong, measured against a spec in 03 and against an issue report in 04. Both say the interviewers looked at how the AI assistant was used, not only at the fix: in 03 to read the spec and the code, in 04 to reach and check the fix.
Keep practicing
- Bloomberg project-coding questions: Bloomberg codebase exercises with an AI assistant (sign-in needed).
- 2025–2026 Bloomberg New Grad Interviews: 98 Experiences (Part 1): the new grad cohort, covered separately, from before these AI rounds showed up in reports.
- Bloomberg interview guide: an overview of how Bloomberg interviews.
- Bloomberg hiring insights: hiring data for Bloomberg.
- Interview experiences: the full feed of reports across companies.
AI coding rounds at other companies
- Meta: Meta AI coding interview: two onsite rounds where you finish a hard algorithm inside a mostly written codebase with an AI model, then face stress-test suites.
- Google: Google AI coding interview: an online assessment with HackerRank's built-in assistant, and a live round on a multi-file project where using AI was optional.
- Stripe: Stripe AI coding interview: experienced engineers kept getting the same transaction rule parser, built step by step with a built-in model.
- Stripe interns: Stripe intern AI coding round: ten intern accounts of the same AI-assisted round.
- Rippling: Rippling AI coding interview: HackerRank's built-in AI on Rippling's familiar problems, plus an HR codebase at the onsite.
- DoorDash: DoorDash AI coding interview: AI CodeCraft: build and run several small services live with your own AI tool.
- LinkedIn: LinkedIn AI coding interview: one long, multi-part task in an IDE whose agent edits files directly.
- Shopify: Shopify AI coding interview: pair programming with your own AI on Shopify's well-known question set.
- Datadog: Datadog AI coding interview: bring any AI tool and share your screen; the tasks are easy, and the follow-ups decide the round.
- Capital One: Capital One AI coding interview: new grad Power Day agentic round: fix a bug, then add a feature to a large full-stack app with Claude Code.
- xAI: xAI AI coding interview: new grad take-homes and onsites where you build with AI.
- Adobe: Adobe AI coding interview: AI showed up as one step in a director round and in a take-home; the rest stayed classic.
- Snowflake: Snowflake AI coding interview: design and build a rate limiter or a retrying task queue while the AI writes the code; one AI Engineer onsite was in person with your own laptop.
- Snap: Snap AI coding interview: a backend-only chat service from an empty Python project with Claude or Codex, or a feature in an existing codebase.
- Ramp: Ramp AI coding interview: build inside a repo Ramp provides; one repo logged every prompt for the interviewer to read.
- Anthropic: Anthropic AI coding interview: review four pull requests, then improve one with Claude Code.
- Apple: Apple AI coding interview: CoderPad with a sidebar agent, alongside Apple's other recent interviews.
More companies with AI rounds on FastPrep:
- Amazon: Amazon interview report: an AI-assisted debugging OA (1 experience).
- Atlassian: Atlassian interview report: project coding and AI tool use (11 experiences, two parts).
- Circle: Circle interview report: three AI rounds: building, debugging and PR review (1 experience).
- Volta: Volta interview report: a deliberately vague data task with a token budget (1 experience).
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.
Comments
0No comments yet — be the first to share what you know.