Design and implement an LRU Cache that supports get(key) and put(key, value) in O(1) time.

Start by naming the two required data structures: a hash map for O(1) lookup and a doubly linked list to track recency order. Draw the structure before coding. Show the move-to-front on get and put, and the evict-from-tail on capacity overflow. Write the full implementation with test cases as requested by at least one reported Apple interviewer. Clean abstraction, handling of the dummy head/tail sentinel nodes, and discipline to actually write working code rather than pseudocode.

Design a HashMap from scratch — implement put, get, and remove without using built-in hash map libraries. Discuss collision handling.

Pick a fixed-size array of buckets with chaining (linked list per bucket). Walk through hash function choice, load factor concerns, and when to resize. Apple interviewers specifically ask about the cycle count (big-O) for each operation. Understanding of hash internals beyond the API, including collision resolution strategies and amortized analysis on resize.

Merge K Sorted Lists — merge k sorted linked lists and return it as one sorted list.

Introduce a min-heap of size k seeded with the head of each list. Each pop yields the smallest node; push that node's next if it exists. Compare with the divide-and-conquer merge approach for the same O(N log k) complexity. Discuss why a naive k-way linear scan is O(Nk). Recognizing the heap pattern without a hint, clean linked-list pointer manipulation, and proactive complexity comparison.

Design iCloud Photos sync — how do you sync a user's photo library across all their Apple devices reliably and efficiently?

Clarify scale (billions of assets, billions of devices). Identify the three hard sub-problems: change detection (diff/hash), conflict resolution (last-write-wins vs user arbitration), and bandwidth efficiency (delta sync, deduplication by content hash). Propose a chunked upload/download pipeline, a metadata service separate from asset storage, and a sync-state machine per device. Explicitly address on-device vs server-side encryption and what the server never sees. Handling of multi-device eventual consistency; proactive discussion of Apple's privacy model; practical bandwidth constraints (cellular vs Wi-Fi tiering).

Design iMessage — a real-time messaging system that delivers messages across iPhone, iPad, Mac, and Apple Watch, including delivery receipts and read receipts.

Model the server as an encrypted courier: it routes opaque payloads and never holds plaintext. Design the delivery pipeline (push via APNs for online, store-and-forward for offline), per-device message ordering, and deduplication. Discuss end-to-end encryption at the architecture level (each device has a public key registered in the directory; senders encrypt per-device). Receipts are lightweight ACK messages through the same pipeline. Understanding that Apple's privacy stance shapes architecture decisions; multi-device fanout; offline delivery; ordering guarantees at scale.

Design Apple Push Notification Service (APNs) — a system that delivers hundreds of millions of push notifications per day to iOS/macOS/watchOS devices worldwide.

Separate device registration (persistent connection from device to APNs gateway, stored token mapping) from notification dispatch (publisher pushes to APNs, APNs routes to the right gateway, gateway writes to device's persistent TCP/TLS connection). Cover connection persistence, reconnect handling, per-app rate limiting, and what happens when a device is offline (TTL, replacement of non-urgent messages). The long-lived connection model vs stateless HTTP; global routing at scale; graceful handling of device offline/reconnect without flooding.

Tell me about a time you disagreed with your manager or a senior stakeholder. What did you do, and what was the outcome?

Use STAR and lead with the data or user research you gathered before the conversation, not just your opinion. Show you came prepared (mockups, metrics, user-testing clips), proposed options with tradeoffs, and either changed their mind or accepted their call with full commitment. Apple specifically values intellectual honesty paired with collaborative execution. Ability to hold a principled position without damaging relationships; data-driven persuasion; no lingering resentment if overruled.

Tell me about a time you completely failed. How did you recover?

Name a real failure with measurable impact — not a reframed success. Describe how you diagnosed the root cause, what you changed in your process, and concrete evidence the fix worked. Apple interviewers probe follow-up questions; you need a story you can go three levels deep on. Self-awareness and ownership without excessive self-flagellation; a credible learning loop; resilience demonstrated by the follow-up.

What were some of your best and worst days at work over the last few years?

This is an Apple-specific question reported by multiple candidates. For the best day, tie it to meaningful impact or craft — shipping something users loved, a hard bug cracked, a junior engineer's breakthrough. For the worst day, pick something professional (not personal hardship) and pivot quickly to what you learned or changed. Authentic motivation — Apple is explicitly checking whether you are intrinsically driven by the work itself, not titles or compensation.

How would you deploy a large language model on-device (e.g., for Siri or Apple Intelligence) given the memory and latency constraints of an iPhone?

Cover model compression techniques in order of aggressiveness: quantization (INT8/INT4), pruning, knowledge distillation. Discuss how Apple Silicon's unified memory architecture changes the tradeoff vs CPU+GPU split. Propose a hybrid: heavy reasoning in the cloud, latency-sensitive or private inference on-device. Address the privacy boundary — what never leaves the device and why. Concrete knowledge of quantization tradeoffs, not just naming them; awareness of Apple Silicon specifics; on-device vs cloud split rationale grounded in privacy, not just latency.

Apple Interview Questions

Apple's SWE loop (ICT2–ICT5) runs 5–7 rounds: a recruiter screen, one or two phone/video coding rounds, and an onsite with 2–4 coding sessions, 1 system-design session (ICT3+), and 1–2 behavioral/hiring-manager rounds. Apple weighs behavioral fit more heavily than most product companies, and system-design prompts are almost always Apple-ecosystem-anchored (iCloud, iMessage, on-device privacy).

17 real questions across 4 rounds

Reported via Glassdoor Apple SWE interview reviews 2024–2025; LeetCode Apple company tag community aggregators; GeeksforGeeks Apple on-campus 2024 report; onsites.fyi ICT3/ICT4 2025 guide; interviewpilot.dev ICT3/ICT4 2025–2026

4 more coding questions Apple actually asks, each with how to approach it and exactly what the interviewer is evaluating. Unlock the full set across every round.

Continue with Google

Or run a Apple mock with the AI interviewer. It asks questions like these, follows up on your answers, and tells you exactly what to fix. The only score that means anything is the one on your real answers.

Continue with Google

Keep prepping

All company interviews