Whiteboard Confidential

A candidate works through three algorithmic problems in Go during a mock interview with a senior FAANG engineer. This mock touched on correctness, complexity tradeoffs, implementation clarity, and communication — the core signals real interviewers evaluate.🧩 Problem 1 — Longest Consecutive Sequence (Unsorted Array)Design an O(n) solution using a hash set to identify sequence starts and expand. Covers duplicate handling, boundary detection, and correctness reasoning without sorting.🧩 Problem 2 — Binary Tree Level Order TraversalClassic BFS over tree nodes, using a queue and level-by-level batching. Discussion includes space complexity, queue growth, and common implementation pitfalls.🧩 Problem 3 — Cutting Trees on a Cliff (Grid Stability / Connectivity)After removing one tree, determine which remaining trees are still connected to the bottom row. Involves multi-source BFS/DFS from stable nodes, component tracking, and correctness guarantees for 4-directional connectivity.
Key skills demonstrated• Translating verbal reasoning into clean Go implementations• Maintaining communication while coding (a major FAANG evaluation axis)• Applying BFS/DFS patterns across trees and grids• Identifying optimal vs. suboptimal approaches via time/space analysis• Handling interviewer hints and recalibrating solutions• Ensuring correctness through example-driven verification👉 Book coaching or watch more mock interviews: https://www.interviewing.io📝 Interview transcript & feedback: https://interviewing.io/mocks/faang-go-longest-consecutive-sequence🔗 Explore more FAANG interviews: https://www.interviewing.io/mocksDisclaimer: All interviews are shared with explicit permission from the interviewer and interviewee. All candidates remain anonymous.

REPLAY EPISODE: A candidate takes on two Meta-style interview problems—one array, one binary tree—and a Senior Meta engineer breaks down their performance step by step.In this Meta-style mock coding interview, a candidate is put through two challenging problems that test practical reasoning, optimal data structures, and clear communication — exactly what real Meta interviewers look for.🧩 Problem 1: Given a sorted array of positive integers and a target k, find the pair whose sum is closest to k.(A classic Meta-style twist on two-pointer optimization problems.)🧩 Problem 2: You’re given the root of a binary tree. If one person stands on the left side and another on the right side, return the nodes visible to both of them — left view (bottom → top) followed by right view (top → bottom).(A unique combination of tree view and traversal-order reasoning.)Watch how the candidate breaks down both problems like in a real Meta onsite:✅ Identifies the optimal O(n) two-pointer strategy✅ Handles tricky edge cases with clean logic✅ Designs a combined traversal strategy for dual tree views✅ Codes efficiently under pressure✅ Gets detailed feedback from a Senior Meta engineer💡 What you’ll learn:• Two-pointer patterns Meta loves to ask• How left/right tree views work and how to combine them• How to structure reasoning during a high-pressure interview• What great communication looks like in a real FAANG interview👉 Sign up to book coaching or watch more mock interviews: https://www.interviewing.io📝 Interviewer feedback & transcript: https://interviewing.io/mocks/facebook-python-closest-pair-sum🔗 Explore more Meta interviews: https://www.interviewing.io/mocks?company=metaDisclaimer: All interviews are shared with explicit permission from the interviewer and the interviewee. All candidates remain anonymous.

REPLAY EPISODE: In this Google machine learning system design interview mock, a candidate tackles a personalized newsfeed recommendation system — the kind of large-scale ML challenge that real Google engineers face.

🧩 Problem: Design an ML system that ranks and recommends posts in a user’s feed by predicting engagement (likes, comments, shares) in real time.

Watch how the candidate approaches it like a real interview:
✅ Clarifies goals, scope, and constraints for a production ML system
✅ Defines the ML objective and key features (user, content, interaction)
✅ Chooses and explains a two-tower deep learning architecture with multitask learning
✅ Discusses tradeoffs in retrieval, ranking, latency, and scalability

💡 What you’ll learn:
• How to approach ML system design questions in Google interviews
• How to connect engagement metrics to ML objectives
• What a two-tower recommendation model looks like in production
• How top candidates communicate complex ML ideas clearly

👉 Watch more interviews or book ML interview coaching: https://www.interviewing.io

📝 See the full transcript and interviewer feedback:https://interviewing.io/mocks/google-machine-learning-personalized-newsfeed-system

🔗 More Google interviews:https://interviewing.io/mocks?company=googleDisclaimer: All interviews are shared with explicit permission from both the interviewer and the interviewee, and all interviews are anonymous. interviewing.io has the sole right to distribute this content.

REPLAY EPISODE: In this Google coding interview mock, a candidate tackles one of the trickiest binary tree problems that keeps showing up in real interviews.

🧩 Problem: Given the root of a binary tree and a list of node values to delete, remove those nodes and return the roots of all remaining subtrees. (LeetCode #1110 – Delete Nodes and Return Forest)
Watch how the candidate breaks it down from scratch:
✅ Clarifies the problem and edge cases like a real Google interview
✅ Designs an O(n) BFS solution (with clean logic)
✅ Codes in Python and tests multiple tricky examples
✅ Gets detailed interviewer feedback on problem-solving, communication, and structure
💡 What you’ll learn:
• How to think out loud in a Google interview
• BFS vs DFS strategies for tree deletion problems
• How to handle “delete and return forest”–type graph problems
• What great communication looks like under pressure

👉 Sign up to book coaching or to watch more interviews in our showcase: https://www.interviewing.io
🔗 Or view other Google interviews: https://interviewing.io/mocks?company=googleDisclaimer: All interviews are shared with explicit permission from the interviewer and the interviewee, and all interviews are anonymous. interviewing.io has the sole right to distribute this content.

SPOTLIGHT EPISODE: What happens when a senior engineer stops “faking it” and starts treating interviewers like partners? Eamonn sits down with James—now at OpenAI, with prior stops at Pinterest, Reddit, and Discord—to unpack how he managed imposter syndrome, learned to ask clarifying questions early, and used improv to become a clearer, more confident communicator on the job.
Along the way, we pause to analyze James’s mock coding interview on interviewing.io, tracing the exact habits that set him apart: writing while thinking, checking in for buy-in, and iterating toward better solutions under pressure. If you’re preparing for coding interviews, you’ll see how James navigates anagrams with letter-count tuples, weighs time/space tradeoffs, and engages with hints without losing the thread.
Watch James's Full Mock Here: https://start.interviewing.io/interview/Gw58xSkxBKqu/replayWatch Another of James's Full Mocks Here: https://www.youtube.com/watch?v=eHF6YxfOXy4
Watch replays & transcripts: https://start.interviewing.io/showcase
Book a mock interview: https://interviewing.io

Timestamps
00:00 Cold open — why “faking it” fails in interviews
01:00 James’s background and path to OpenAI
04:00 How improv rebuilt confidence and communication skills
07:00 Layoffs at Discord and finding upside in a reset
10:00 Structured prep: LeetCode, mocks, and recruiter advice
13:00 Why peer mocks matter more than grinding alone
16:00 Mock interview replay: the anagram challenge
24:00 Working through space vs. time tradeoffs under pressure
33:00 Staying unstuck: writing, asking, and engaging hints
58:57 Feedback clip: top-tier insight, senior-level signal

Disclaimer: All interviews are shared with explicit permission from the interviewer and the interviewee, and all interviews are anonymous. interviewing.io has the sole right to distribute this content.