What Is the Mercury Interview Process Like? (Round by Round)
Candidates report four stages at Mercury, the business banking company. The process starts with a recruiter screen and continues with a technical screen with one engineer, where backend candidates report a SQL pull request review and frontend candidates report React and TypeScript questions instead.
Then comes a longer technical session, about three hours in one report, with independent coding and a shared code review. The last stage is a set of conversations with the team, including behavioral questions.
Reported end to end times range from about two weeks to about five weeks. Mercury does not publish an official round list, so details vary by role and level.
Quick Overview
| Stage | Format | What is evaluated |
|---|---|---|
| 1. Recruiter screen | Short call | Background, motivation, team preference |
| 2. Technical screen | Live, one engineer | SQL and schema review, or React and TypeScript questions |
| 3. Build and review session | Live, about three hours in one report | Building a small feature and defending its design |
| 4. Team loop | Several conversations | Design, product thinking, behavioral fit |
Stage 1: Recruiter Screen
Expect questions about your background, your interest in banking software, and which product group fits you. Mercury's job posts list product groups such as Banking, Everyday Experiences, Expansion, and Risk. One candidate warns that a weak recruiter screen leaves a poor impression, so prepare questions about the business.
The motivation question appears here first. Prepare it with How to answer why Mercury.
Stage 2: Technical Screen
The backend version is a SQL pull request review. A pull request is a proposed code change that a teammate reviews before it is merged. Candidates report that you review and refactor the database schema in the pull request, so the round tests SQL knowledge and code review skill together. A weak review points only at syntax, while a strong review covers table design, indexes, performance under load, and what happens when the change is only half applied.
The frontend version covers React and TypeScript. Candidates report questions such as what a ref is and why not to overuse refs, how to manage state and why not to use Redux, and whether TypeScript runs in the browser. Either way, speak your reasoning out loud as you go.
Stage 3: Build and Review Session
Candidates report a second technical stage with a much wider scope than the screen. One candidate describes about three hours: an engineer explains the task for 15 to 30 minutes, you code alone for about two hours, and then you review the code together for about 40 minutes. Another candidate describes writing a short spec and then using AI tools to build a small working version.
Mercury publishes guidelines on AI use in its interviews, so use AI in a round only when the recruiter says it is allowed. Published interview guides also describe a separate Craft round about types, invariants, and error handling. Mercury does not publish a round list, so treat the Craft round as a description from guides, not a confirmed stage.
Whatever the exact format, the grading is similar: clean structure, good names, correct handling of edge cases, and a data model you can defend matter more than a clever algorithm. Mercury's backend is written in Haskell, and published guides say you may use any mainstream language in these rounds, so confirm the language with the recruiter.
Stage 4: Team Loop
Candidates report behavioral conversations with the team after the two technical stages. One candidate describes a final two hour session with a presentation and STAR style questions, where STAR stands for Situation, Task, Action, Result.
Published guides describe a system design round with banking topics for some roles. Candidates do not report one for every role, so ask the recruiter whether your loop includes it. The design topics come from Mercury's product: money movement, ledgers, and risk. See What to expect in the Mercury system design interview for the topic list.
Mercury publishes six values: Think actively, Be super helpful, Act with humility, Appreciate quality, Focus on the outcome, and Build relentlessly. Behavioral questions test for these directly, so prepare one story for each.
Timeline and Decision
Reported timelines vary: candidates report an average of about two weeks from application to hire, while published interview guides give three to five weeks from the first recruiter call to an offer.
The reported stages are separate meetings on separate days, not one long onsite day. Candidates rate the difficulty near the middle of the scale. One repeated complaint is a rejection with no explanation after several rounds, and one repeated praise is a process that resembles real engineering work.
For wait times after each stage, read How long does it take to hear back after a Mercury interview.
Questions to Ask Them
Every round ends with time for your questions. Good questions here: how the Haskell training program works for engineers new to the language, how a product group decides what to build next, and how on call works for a system that moves money. Mercury describes its culture as remote first with optional offices, so ask how the team communicates across time zones. A specific question also proves you researched the product.
How to Prepare
- Practice reading code, not only writing it. Review real pull requests and say what you would change, because the SQL screen rewards this habit.
- Drill the fundamentals. Grokking the Coding Interview teaches the patterns you need in the coding session.
- Learn the design of money systems. Grokking the System Design Interview covers the building blocks used in ledgers and payment flows.
- Prepare six value stories. One story per published value, each with a number in the result.

GET YOUR FREE
Coding Questions Catalog

$99

$197

$72