2015–2026 Apple SWE New Grad & Early-Career OA / Interviews: 42 Experiences (Part 3 :)

Candidate experiences 28–42

28: Two US interviews without a conventional LeetCode question

Report: Apr 20, 2022 · Role: New Grad Software Engineer · Location: US

First conversation

A senior engineer introduced the organization and answered the candidate’s questions.

Technical lead

The next interviewer assigned a role-specific coding task and asked domain questions. The candidate explicitly reported no conventional LeetCode question and no behavioral questions. The task’s domain and behavior were not disclosed. More specific kernel, OS, and bit-operation claims appeared in anonymous replies whose authorship was not established as the original candidate’s, so they are not added here.

29: Three phone interviews before a planned six-hour onsite

Report: Jan 18, 2022 · Role: SWE New Grad

Phone interviews

The candidate reported about three 45-minute calls, easy Apple-tagged questions including reversing a list and an unspecified palindrome task, and detailed résumé questions. They did not identify what the palindrome problem asked them to return.

Next stage

They had moved forward with one team to an upcoming onsite lasting about six hours, with 45-minute interviews and 15–30-minute breaks. That schedule was planned at the time of the report, not a completed onsite.

30: A campus SDE1 application with two conflicting write-ups

Interview: Nov 2021 · Role: SDE1, campus · Cohort: One write-up says Fresher, another Experienced · Result: Rejected

Write-up A

The campus applicant described a 60-minute OA with nine MCQs and two named problems: Gas Station, asking for the first starting pump that completes a circular tour or a finding that none works, and House Robber II. The accompanying Gas Station approach was greedy. The linked House Robber statement had nonnegative amounts and a circular street with adjacent first/last houses; its accompanying tree-DP explanation conflicted with that circular task.

A 60-minute technical interview covered car-rental LLD, database normalization, judging whether a relation was in Boyce–Codd normal form (BCNF), and the résumé. A 30-minute manager interview included the two-eggs/100-floors problem; its accompanying approach began at floor 14 and tested the lower floors one by one if the egg broke. The manager also asked how to make the candidate’s social-site or exam-portal project scale by choosing the right database. A 20-minute HR interview asked Why Apple, future plans, and what they knew about Apple. It also asked how many OA problems they had solved and what those problems asked for.

Write-up B

The same author, company, and month appeared in another version: a 53-minute OA with nine MCQs and two unnamed tasks; a 60-minute technical interview with three or four tasks; then a 40-minute project-focused manager conversation.

The disclosed technical tasks included a data structure scheduling processes by priority and arrival time, the middle of a linked list, and locating an original tree node’s counterpart in a cloned tree. The platform’s linked middle-node statement chose the second middle for even lengths and returned None for an empty list. The middle-list approach started with fast/slow pointers, then unexpectedly described inserting a node; that insertion wording did not match the linked middle-node task. Its cloned-tree statement required an unchanged original and copied tree, distinct original nodes, a non-null target, and returning the corresponding node reference.

Full-stack work and deployment also came up. Another question asked them to state the steps in system design and the software lifecycle. These two versions count as one candidate account; their different timings and question lists are preserved rather than combined into a supposedly fixed loop. Expanded platform practice statements supply more detail than the candidate’s short recollection.

31: One graduate interviewed with three Apple teams

Interview period: Jul–Oct 2021 · Role: Graduate SWE search · Graduation: Dec 2021

After an Apple internship

The candidate’s internship did not convert because of headcount, so they applied internally to several teams. The article title referred to 2022 new grads; these interviews happened in 2021.

AMP Commerce

The September 7 phone interview included lowest common ancestor. A second phone interview on September 14 used an unnamed Trie Hard problem. The September 24 virtual onsite had five conversations: database-schema design; Subarray Sum Equals K; an unnamed DFS Hard; an unnamed DP Hard; and hiring-manager/behavioral discussion. They were first told they had passed, then learned there was no headcount.

AI/ML Rotation

An information call on July 27 was followed by an August 17 phone interview with an unnamed stack Medium task. The candidate was rejected.

ML Infrastructure

The September 14 phone interview had two unnamed easy questions. The September 28 onsite included two Python coding interviews with easy tasks, ML-platform system design, and three behavioral/hiring-manager conversations. The hiring-manager conversation also included an unnamed brain-teaser. The candidate was rejected here too.

The account covers one graduate’s search across three teams. The ML-related teams are identified separately from AMP Commerce, and the unnamed Trie, DFS, DP, stack, and Python tasks are not assigned familiar problem numbers.

32: BITS campus: three named algorithms across the OA and interview

Interview period: Mar–Apr 2021 · Role: SDE1, Fresher · Result: Selected

OA — 60 minutes

The two tasks were Merge Intervals and Clone Graph. The expanded interval statement used closed integer intervals: a shared endpoint counted as overlap, and the output was ordered by start. The graph statement used a connected undirected graph with node values and neighbor lists and asked for a deep copy. Its accompanying approach used a hash map to avoid repeated visits.

Video interview — 60 minutes

Check whether two binary trees have identical structure and values, followed by project discussion.

HR — 45 minutes

Self-introduction and dealing with conflicts in projects. Although the platform counted six “coding problems,” some entries were project or behavioral prompts; the account named three algorithms.

33: Shazam’s twelve-month graduate placement OA

Report: Feb 1, 2021 · Role: Graduate Software Engineering Placement · Term: 12 months

HackerRank OA

The candidate had seven days to begin a 90-minute assessment, with 19 languages available. One task asked for the largest later-array-value minus an earlier value, returning -1 when no positive difference existed. For [3,6,10,1,4,6,5], the result was 7. They recalled a minimum input length of one for that task. The other asked for the length of a strictly increasing subsequence; [1,1,3,2,6,4,1,7] had length 4, for example [1,3,6,7], and a fully descending array returned 1.

This was a fixed-term graduate placement, not an ordinary permanent new-grad role.

34: India: an OA, two runnable coding interviews, and HR

Interview: Nov 2020 · Role: SWE New Grad · Location: India · Background: IIT undergraduate · Result: Offer

OA — 90 minutes

Ten computer-science MCQs and two coding tasks. The coding prompts were not disclosed.

First technical interview — about 45 minutes

On Webex, the candidate solved Number of Islands and Merge Intervals, ran their code, and discussed an internship.

Second technical interview — about 60 minutes

Two interviewers moved from Two Sum to 3Sum, discussing brute force and improved approaches. The report explicitly described implementing 3Sum, running it, finding a bug, and correcting it. Internship and behavioral discussion followed.

HR

Further-study plans, expectations, and what the candidate learned from the internship. An offer followed.

35: A graph problem in a first virtual onsite

Interview: Oct 2020 · Role: New Grad SWE · Candidate: Master’s graduate, zero years of experience · Result: Rejected

Apple interview

In a multi-company job-search account, the candidate described Apple’s first virtual-onsite interview as one graph problem using backtracking/DFS that they had not seen on LeetCode. They wrote and ran a solution, then received a rejection three days later. The exact graph task was not supplied; none of the other companies' questions is part of this Apple account.

36: Platform Architecture: a GHC question followed by a C queue

Interview period: 2019; report Dec 10, 2019 · Role: SWE New Grad · Team: Platform Architecture

GHC conversation

After roughly 15 minutes of OS questions, the candidate was asked to find the first and last positions of a target in a sorted array.

Phone interview

Implement a queue in C and handle special cases. Follow-ups asked how condition variables, a mutex, and a semaphore related to that implementation, then covered OS, computer architecture, and compiler fundamentals. The report did not specify a queue capacity, producer count, or blocking API; the practice exercise makes those choices explicit.

37: AMP: return every element after the target

Report: May 26, 2019; body labels the interview May 30 · Role: New Grad Software Engineer, AMP · Location: Cupertino

CoderPad phone interview — one hour

Given a list of objects and a target element from that list, return every element after it. For [1,2,3,4] and target 2, the answer was [3,4]; a final-element target returned an empty list, and a null target was an error. The interviewer followed up on efficiency and complexity.

Another phone interview and an onsite were planned. The report’s body date and publication date conflict, so the interview is placed in May 2019 without selecting an invented exact day.

38: An AMCAT OA followed by arrays, pairs, coins, and palindromes

Interview period: Mar–Apr 2019 · Role: SDE1, campus source · Cohort: Platform labeled Experienced

AMCAT OA — about 53 minutes

Nine MCQs and two Medium coding questions; one was explicitly Merge Intervals. This is direct platform evidence for this assessment, not for every Apple OA.

Technical interview — 60 minutes

Find the shortest contiguous subarray whose sum is at least K, returning -1 if none exists. Count index pairs whose values sum to a multiple of K; repeated values at different positions can make different pairs.

Another technical/behavioral interview — 60 minutes

Count combinations using unlimited coins of values 3, 5, and 10, with order irrelevant. Find the nearest different palindrome by absolute distance, choosing the smaller one on a tie; the expanded example mapped 121 to 111.

The campus setting and the platform’s Experienced label are both retained. This account is adjacent to the confirmed graduate reports.

39: IIT campus: Josephus, a leaderboard, and DNS

Report: Jan 22, 2019 · Role: Campus software engineering · Interview date: Not supplied

Three interviews, 50 minutes each

The coding interview asked for an analytical Josephus solution and an unnamed DFS problem. The interviewer expected runnable code, useful function names, and correct boundary handling.

The design interview asked for a game leaderboard showing top-ranked players by region. Follow-ups covered databases, SQL versus NoSQL, APIs, reliability, availability, latency, and the CAP trade-off. The candidate was also asked what happens after a browser receives a URL, how DNS works, and how to design DNS.

HR explored projects, internship scalability, open-source contributions, and working culture. The source was published in 2019; a later 2024 update timestamp does not move the interview into 2024.

40: Cupertino: two phone interviews, with a next-leaf question

Interview: May 2018 · Role: SWE New Grad · Location: Cupertino

First phone interview — 45 minutes

The opening 15–20 minutes covered introductions and the résumé. An unnamed string-based LeetCode Hard followed.

Second phone interview — 45 minutes

Given a leaf node in a tree, find the next leaf. The recollection did not fully define the tree interface or traversal convention; the linked practice states its completion explicitly.

41: Boston: implement atoi in C

Interview: Oct 2015 · Role: SWE New Grad · Location: Boston

Technical phone interview — 45 minutes

A recruiter arranged the call. The first 15 minutes covered the résumé and projects; the coding task was to implement atoi in C. The candidate mainly used Java and Swift and had not prepared C. The recollection did not provide exact whitespace, invalid-character, or overflow rules; those are stated separately in the practice version.

42: A campus interview of CS and Java questions

Interview: Sep 2015 · Role: SWE New Grad · Channel: Campus recruiting

Technical discussion — 30 minutes

There was no coding implementation. Questions covered threads versus processes, deadlocks, hash maps and their complexity, basic algorithms, and differences between Java classes and implementations. The report did not disclose specific Java class names or separate runnable tasks.

In this series: Part 1 · Part 2 · Part 3 (current)

Sources: Anywhere useful interview info can be found.