2025–2026 ZipHQ New Grad OA / Interviews: What Candidates Shared

A roundup of public ZipHQ new grad interview reports, with a few other SWE reports where they add useful detail.

Coverage: New grad OA, phone screens, onsite rounds, and project discussions reported from March 2025 to April 2026. One NG report names San Francisco; the others do not give a location. Other SWE reports from 2025–2026 are labeled separately below.

The short version

The new grad reports include both algorithm questions and practical coding.

1. Prepare for more than LeetCode. Candidates described OOP coding, a small implementation task, or adding a feature to an existing codebase.
2. Leave time for follow-ups. Difficulty estimates ranged from Medium to Hard. One candidate mentioned a graph question with a harder follow-up.
3. Know your projects well. A project discussion appeared in several accounts. One candidate spent a full hour talking through a specific project.

A useful prep mix: timed algorithms, a small coding task in an unfamiliar project, and a clear walkthrough of something you built.

Online assessments

OA 01: What the new grad reports actually say

The candidate who received an offer after interviewing in November 2025 described several standard Easy/Medium questions. Another NG candidate named CodeSignal as the platform.

Neither shared enough detail to reconstruct the assessment. The reports do not establish a fixed question count, time limit, or passing score for NG applicants.

Reported topics and tasks

Algorithms

• Trees with memoization.
• Priority queues and arrays.
• An optimal-path problem on a graph, followed by a harder question.

These are the details the NG sources gave. They did not share complete prompts or reliable LeetCode numbers.

Practical coding

The NG accounts mention an OOP-style round, implementing a small task, and adding a feature to an existing codebase. They do not explain the exact feature or business rules.

Projects and behavioral questions

One candidate was asked to explain a specific project in depth. Other reports mention a behavioral round or a hiring-manager conversation. The few disclosed questions include introducing yourself and explaining why you want to join Zip.

Five new grad reports

These are three Glassdoor reports and replies in two Reddit threads. The reports are anonymous, so it is not possible to rule out overlap between sites.

01: An offer after a two-week process

Reported: December 4, 2025 · Interviewed: November 2025 · Result: Accepted offer

The candidate applied online. After the OA, there was a recruiter screen and a one-hour algorithm interview.

The next stage had three separate one-hour sessions: another algorithm interview, practical coding, and a detailed project discussion. The coding questions were described as Medium/Hard. The practical task was a small implementation exercise; no prompt was shared.

The candidate found the interviewers friendly and professional. The recruiter also explained the offer clearly.

02: Two algorithm rounds, plus OOP and a project discussion

Reported: December 22, 2025, with a follow-up in April 2026

A reply in a new grad thread described one Medium question in the first technical screen. The virtual onsite included another Medium question, OOP-style coding, and a behavioral/project discussion.

The same commenter said one algorithm question used trees and memoization, while the other involved a priority queue and arrays. In the April reply, they identified trees as the first-screen topic. No outcome or full question statement was shared.

03: A graph question and a feature in an existing codebase

Reported: November 14, 2025 · Location: San Francisco · Result: No offer

This NG candidate described CodeSignal OA, a recruiter call, a round they called system design, and a three-round superday with two technical interviews and one behavioral interview.

One technical task involved implementing a feature in an existing codebase. The disclosed algorithm topic was an optimal graph path with a harder follow-up. The candidate rated the coding as mostly Hard. For practice, try the related Shortest Distance in a Graph problem.

Only this NG report mentions a separate system-design round, so it may not be part of every candidate’s process.

04: The manager chat depended on the technical results

Reported: January 14, 2026

Another reply described an initial technical screen, followed by two technical rounds back to back. Candidates who passed then moved to a hiring-manager chat. Otherwise, that chat was canceled.

The reply did not include question details or an outcome.

05: The role was filled before the process went further

Reported: March 9, 2025 · Result: No offer

This candidate initially felt the process was clear, then stopped hearing back. After following up, they learned the role had been filled.

The only disclosed question was why they wanted to join Zip. There is no technical question to take from this report.

Other SWE reports worth reading alongside the NG accounts

These reports add context, but none confirms the current new grad process.

Working in multiple files

A San Francisco SWE candidate who interviewed in June 2026 described a recruiter call followed by pair programming. They had prepared mainly with LeetCode and felt unprepared for a project spread across multiple files. The role was not labeled NG. The question field mentions two tasks “with the file system,” but gives no clear problem statement.

An October 2025 Senior SWE report was more specific: two one-hour screens, one algorithmic and one practical, both done in Python. The practical round used about ten Python files and required a feature implementation plus unit tests. Requirements became more complex as the interview progressed. Treat the language and testing requirements as details of that senior interview.

CodeSignal OA details

A January 15, 2026 SWE report named the CodeSignal Industry Coding Assessment. The candidate said all tests passed but reported no follow-up.

In a separate January 27 report, the recruiter reportedly said passing “3 out of 4” would lead to the next round. The candidate said they nearly completed all four but were rejected. The report calls the assessment OOD, but does not say whether “four” means separate questions or progressive levels. Neither account establishes an NG cutoff.

A more detailed graph question from Toronto

A Toronto SWE candidate who accepted an offer in January 2026 reported a shortest-path problem on an infinite grid. It included ordinary barriers and moving barriers with a head that blocked a row or column before and including that head. The report did not explain the movement rules well enough to recreate the problem. For a practice version with the movement rules spelled out, try Shortest Path Around Moving Parades.

Their process included a recruiter call, an algorithm screen, onsite algorithm and practical/OOD sessions, and a manager discussion about projects or internships. The post did not identify the role as NG.

AI rules can differ between rounds

A September 2026 full-stack SWE report described a full-stack practical round on CodeSignal that they called “no AI.” The onsite then included system design, AI coding, regular coding, and a project discussion.

That is a report about a full-stack role. It does not show that NG candidates get the same rounds. Check the rules for each interview before preparing around an AI tool.

What I would prepare first

Start with tree recursion and memoization, heaps, arrays, and graph traversal. Practice explaining your approach while coding, then leave time to check edge cases and discuss a follow-up.

For practical coding, try working in a small project you did not write. Find the relevant code, make one change, and check that it works. Writing a few tests is useful practice, even though the NG reports do not specify a testing requirement.

For the project conversation, choose something you know beyond the README. Be ready to explain your own work, a decision you made, a bug you fixed, and what you would change now.

Final checklist

• Practice a timed algorithm question with a follow-up.
• Make a small change in an unfamiliar codebase.
• Walk through one project without relying on a memorized script.
• Confirm the next round's format, allowed languages, and AI rules.