Search the siteSearch in

Interview paths
Question bank
Coding practice
Study materials
Infrastructure lab

Community

Weekly sessions
Peer practice
Contribute

News

Jobs

Open roles
Market intelligence
Companies
Roles

Feedback

Product feedback
Developer feedback

Checking account…

Home
Question bank

Frontier AI technical interview question bank.

Practice coding, system design, behavioral, and project questions for frontier AI roles.

Company

Loading the question catalog…

Preparing this page…

Your current navigation and account state stay available while the next page loads.

Frontier Rehearsal · Independent practice · Not affiliated with OpenAI, Anthropic, or xAI.

Home Interview questions Company interview processes Company reports Role reports Market intelligence Open roles Python lab Materials How it works About Feedback Uploads Feature requests Terms Privacy 中文 English

Home
Question bank

Frontier AI technical interview question bank.

Practice coding, system design, behavioral, and project questions for frontier AI roles.

Company

FREE PRACTICE

Public demos, always pinned.

Start with either demo. They remain here when you filter or sort.

Free account: sign in to practice the marked Free questions.

Newest published questions; deterministic, never randomized.

Demo 1Number search with delayed feedback
Demo 2Design real-time device monitoring and control

Question bank filters

Reset filters

Question groups

Filter question bank by interview groupAll 269 Coding 162 System design 48 Behavioral 29 Hiring manager 16 Project deep dive 14

Topics

Filter question bank by topic tagsAll 269 Python 161 Algorithms 40 Data structures 46 Concurrency 27 Distributed systems 7 Storage & state 43 Networking 7 Reliability 30 ML / AI 27 Data / SQL 24 Security 7 Product design 25 Communication 66

Independent and not affiliated with the companies represented. Collected material does not predict future interview questions.

215 questions

Sort question bankSortFree first Newest Oldest AI ratio ↓AI ratio ↑

Question bank results

Every question has a public brief with the task, constraints, and follow-ups. Labels describe interactive practice access.

Public brief

Topic tags
Python
Data structures

xAICoding

Python codingWebAssembly + public tests

Reported interview periodFeb 2026

Weighted LRU Cache

Build an LRU cache whose capacity is the total weight of stored entries rather than the number of keys.

Practice focusget returns -1 when absent and marks a hit most-recently used. put inserts or replaces a key. On replacement, remove the old weight before checking capacity. Evict least-recently used entries until the new value fits; one insertion may evict several entries. A non-positive size or a size exceeding capacity raises ValueError without changing the cache. Exact fits are valid. Target average O(1) time for get and for each individual insert, update, or eviction. The provided run case owns fixture dispatch and is not candidate work.…

Public brief

Topic tags
Python
Data / SQL

xAICoding

Python codingWebAssembly + public tests

Reported interview periodJan 2026

Live Audio Room Analytics

Events have (operation, room_id, user_id, timestamp), where operation is create, join, or leave. Creating a room also places that user inside it.

Practice focusImplement a batch function that returns total person-seconds per room and a live tracker that ranks rooms by current participant count: For batch processing, sort by timestamp and ignore exact duplicate events. A user may occupy only one room: joining another closes the previous stay at that timestamp. Rejoining the same room is a no-op, and an unmatched leave is ignored. Open stays count through current time when supplied; otherwise they contribute nothing. Live top-k breaks count ties by room ID.…

Public brief

Topic tags
Python
Concurrency

xAICoding

Talk-through codingNo local grader

Reported interview periodDec 2025–Jan 2026

Sort Integers with Multiple Workers

Implement an ascending integer sort that divides sufficiently large inputs into chunks, sorts chunks concurrently, and merges the sorted runs.

Practice focusReturn a new list and do not mutate the input. For fewer than 1000 values or a single worker, a direct sort is acceptable. On the large-input path, actually partition the input and dispatch chunk sorts; do not sort the entire list in one call. Cap the worker count to useful work and merge all runs correctly, including duplicates and negative values. The worker count must be a positive integer (not a boolean); reject invalid counts with ValueError before dispatch. Use native worker execution for the concurrent path.…

Public brief

Topic tags
Python
Reliability

xAICoding

Python codingWebAssembly + public tests

Reported interview periodDec 2025

Identify Faulty Nodes with Group Tests

There are n compute nodes with IDs 0 through n - 1. A supplied function test(group) returns True exactly when every node in group is healthy. A test group must contain at least two node IDs. Assume at least two…

Practice focusImplement: Return every faulty node ID in ascending order. Minimize unnecessary tests, but prioritize correctness. Do not assume that testing one node is legal. The supplied run case owns fault labels and the test callback; implement only find faulty nodes . Each test must contain at least two distinct valid non-boolean integer IDs. The public safety budget is at most n (n-1)//2 + n-1 callback calls; it admits pair discovery followed by classification and is not an optimality claim. Nodes are fixed during a run, and n is an integer at least two.…

New public brief

Topic tags
Python

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Best Time to Buy and Sell Stock

Given daily prices, compute the maximum profit obtainable by buying once and selling once later. You may choose not to trade, in which case the profit is 0.

Practice focusPrices are non-negative integers. An empty or one-day input returns 0 . Solve the problem in O(n) time and O(1) auxiliary space. Staged follow-ups 1. Allow any number of non-overlapping transactions while holding at most one share. 2. Add a fixed transaction fee or a one-day cooldown after selling. 3. Return the chosen buy and sell days with deterministic tie-breaking.

New public brief

Topic tags
Python
Algorithms

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Binary Search

Given a strictly increasing array of integers and a target, return the target's index or -1 when it is absent.

Practice focusThe solution must run in O(log n) time and use O(1) auxiliary space. An empty array returns -1 . Staged follow-ups 1. Return the first position where the target could be inserted while preserving order. 2. Support duplicate values by returning the first matching index. 3. Search a conceptually unbounded sorted data source through a bounded-access interface.

New public brief

Topic tags
Python

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Count Values with an Even Number of Digits

Given a list of positive integers, return how many contain an even number of decimal digits.

Practice focusFor example, [12, 345, 2, 6, 7896] returns 2 . Assume every value is positive and fits in a signed 32-bit integer. Target O(n) time. Staged follow-ups 1. Extend the definition to include zero and negative integers without counting a sign. 2. Avoid converting numbers to strings. 3. Count digits in an arbitrary base supplied at runtime.

New public brief

Topic tags
Python
Storage & state

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Course Schedule

There are num_courses courses labeled 0 through num_courses - 1. Each prerequisite pair (course, prerequisite) states that the prerequisite must be completed first.

Practice focusImplement: Return whether it is possible to complete every course. Courses with no edges are valid, and duplicate prerequisite pairs should not change the answer. Aim for O(V + E) time. Staged follow-ups 1. Return one valid course order, or an empty list when none exists. 2. Return an explicit cycle to explain why the schedule is impossible. 3. Process prerequisite additions incrementally and reject an addition that creates a cycle.

New public brief

Topic tags
Python

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Maximum Non-Adjacent Sum

Each value in amounts is the reward available at one location. You may select any subset, but cannot select two adjacent locations. Return the largest total reward.

Practice focusAssume rewards are non-negative. The empty input returns 0 . Produce an O(n) -time solution and then reduce its auxiliary space to O(1) . Staged follow-ups 1. Return the selected indices, with deterministic tie-breaking. 2. Arrange the locations in a circle, so the first and last are also adjacent. 3. Generalize to a tree where directly connected nodes cannot both be selected.

Public brief

Topic tags
Python
Algorithms

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Palindrome Integer

Determine whether a signed 32-bit integer reads identically from left to right and right to left in decimal notation.

Practice focusNegative numbers are not palindromes because of the sign. A positive number ending in zero is not a palindrome unless the value itself is zero. Staged follow-ups 1. Solve without converting the number to a string. 2. Reverse only half the digits to avoid overflow and unnecessary work. 3. Generalize the check to an arbitrary base from 2 through 36.

Public brief

Topic tags
Communication

xAIBehavioral

Discussion / designResponse prompt

Reported interview periodNot available

A thin check

A story about slowing a launch because a safety check was thin.

Practice focusBehavioral round. The interviewer asks for a time you slowed a launch. Answer in this order: the date range, the check, the person, your sentence, their reply, and the agreement. Then they ask what you would have done if the launch could not move. Give the cut you would have made. Then they ask how you knew the check was enough the second time. Point at the measurement, not a feeling. Stay on one project.

Public brief

Topic tags
Communication

xAIProject deep dive

Discussion / designResponse prompt

Reported interview periodNot available

After launch

Where a design failed after launch, what you measured, and what you changed.

Practice focusProject round. One system, after launch. Start with the design you intended. Then the failure, including who felt it. Then the measurement, including the unit and the window. Then the change. The interviewer asks what you left alone and why. Answer with one thing you refused to change. Then they ask how you would know the failure had returned. Point at the same measurement. Do not switch projects.

Public brief

Topic tags
Communication

xAIBehavioralAI-generated

Discussion / designResponse prompt

Reported interview periodNot available

A call you stopped

One behavioral story about stopping a decision that was moving too fast.

Practice focusBehavioral round. Tell one story about stopping a decision that was moving too fast. What the decision was. Who was in the room. The sentence you used. What they said back. What you agreed to measure before continuing. Stay on that one story.

Public brief

Topic tags
Communication

xAIBehavioralAI-generated

Discussion / designResponse prompt

Reported interview periodNot available

A claim you refused

A culture conversation about a claim you would not publish unchecked.

Practice focusCulture round. Someone wants to publish a claim you have not checked. What the claim is. What evidence is missing. What you would ship instead. Who you would tell if they publish it anyway.

Public brief

Topic tags
Storage & state
Communication

xAIBehavioralHalf AI-generated

Discussion / designResponse prompt

Reported interview periodNot available

Fast feedback and correction

Tell one story where you shipped a bounded first version, learned that an assumption was wrong, and corrected it quickly. State the boundary, signal, correction, and impact without hiding the original mistake.

Practice focusPreparation - Write the user or stakeholder, the constraint, and the decision in one sentence each. - Choose one measurement or observable fact that can support the answer. - Name one alternative and the reason it was not selected. Review checklist 1. Did I state the concrete constraint before describing the solution? 2. Did I separate my own decision and work from the team's work? 3. Did I name evidence or a measurement rather than relying on confidence? 4. Can I answer every follow-up without changing the original facts? Staged follow-ups 1.…

Public brief

Topic tags
Python

MetaCoding

Python codingWebAssembly + public tests

Reported interview periodNot available

Maximum profit from repeated stock trades

You are given a price for each trading day and may complete as many round trips as you like, holding at most one share at a time. Report the largest total gain reachable across the whole window.

Practice focusRequirements - The input has two lines: a day count, then that many space-separated integer prices on the second line. The count matches the number of prices. - A round trip means buying one share on some day and selling it on a later day; positions cannot overlap and short selling is not allowed. - Return the total profit as a plain decimal string. A window with no profitable round trip, including a single day, yields 0 . - Prices are non-negative integers and the day count is at least one.…

New public brief

Topic tags
Python

Scale AICodingAI-generated

Python codingWebAssembly + public tests

Reported interview periodNot available

Reconcile noisy annotator labels

Three annotators label each item and report an integer confidence. Produce the final label by weighted vote and count the items where every annotator agreed.

Practice focus- The first line is the item count n . Each of the next n lines has exactly three <label :<confidence pairs, where the label is A or B and the confidence is a nonnegative integer. - Sum the confidence per label. Return A when its total is higher, B when its total is higher, and ABSTAIN on an exact tie. - After the per-item labels, return one final line agreed=<count , where an item is agreed only when all three raw labels match. - Return one label per item in input order, then the agreement line. Do not append a trailing newline.

New public brief

Topic tags
Python

NVIDIACodingAI-generated

Python codingWebAssembly + public tests

Reported interview periodNot available

Batch inference requests under a capacity budget

A serving path receives requests in order. Pack them greedily into batches that respect both a total-size capacity and a maximum request count, then report the batch count and batch sizes.

Practice focus- The first line has two integers: the batch capacity C and the maximum request count per batch B . The second line is the request count n , and the third line has n request sizes. - Walk the requests in input order. Add the next request to the current batch when the total size stays within C and the count stays within B ; otherwise close the batch and start a new one. - Every request must be placed exactly once, in order; a request equal to C forms its own batch. - Return batches=<count ;sizes=<s1 ,<s2 ,... with the batch totals in order.…

New public brief

Topic tags
Python

CursorCodingAI-generated

Python codingWebAssembly + public tests

Reported interview periodNot available

Apply sequential line edits to a file

An editing agent proposes line-range replacements against a small document. Apply them in order, with each edit seeing the document produced by the previous one, and report the final content.

Practice focus- The first line is the current line count m , followed by the m document lines. The next line is the edit count k , then k edit lines. - Each edit line is <start <end <replacement with 1-based inclusive bounds over the current document. Replace those lines with the single replacement line, preserving spaces inside the replacement. - Apply edits sequentially; later edits index the already-updated document. All edits are valid. - Return the final document lines, then a final line lines=<count . Do not append a trailing newline.

Public brief

Topic tags
Python
Algorithms
ML / AI
Product design

Coding exerciseCodingAI-generated

Practice formatRunnable coding45–70 min

Reported interview periodN/A

Build a token-budget batcher

Implement a small batching primitive, then adapt it to padded cost, bounded output, and production reasoning.

Practice focusPython, Greedy algorithms, Boundary cases, ML batching, Complexity through a runnable coding exercise.

Public brief

Topic tags
Python
Algorithms
Data structures
Product design
Communication

Coding exerciseCodingAI-generated

Practice formatRunnable coding45 min

Reported interview periodN/A

Interval calendar

Reason about half-open intervals through conflicts, room counts, merging, and free slots.

Practice focusPython, Sorting, Intervals, Heaps, Edge cases through a runnable coding exercise.

Public brief

Topic tags
Python

Coding exerciseCodingAI-generated

Practice formatRunnable coding45 min

Reported interview periodN/A

Nested dictionary transforms

Flatten and reconstruct nested data, then support path access and lists.

Practice focusPython, Recursion, Data transformation, JSON-like data through a runnable coding exercise.

Public brief

Topic tags
Python
Storage & state

Coding exerciseCodingAI-generated

Practice formatRunnable coding45 min

Reported interview periodN/A

Command-log store

Evolve a tiny state machine through undo, compaction, and snapshot replay.

Practice focusPython, State machines, Undo logs, Compaction, Persistence through a runnable coding exercise.

Public brief

Topic tags
Python
Data structures
Reliability
Data / SQL

Coding exerciseCodingAI-generated

Practice formatRunnable coding50 min

Reported interview periodN/A

Dependency-ready scheduler

Turn a dependency graph into stable ready sets, execution layers, and failure impact.

Practice focusPython, Graphs, Stable ordering, Cycle detection, Failure analysis through a runnable coding exercise.

Previous8 of 9Next

Browse all original practice

Build a token-budget batcher
Sliding-window rate limiter
Interval calendar
Nested dictionary transforms
Command-log store
Dependency-ready scheduler
Streaming metric window
Evaluation slice aggregator
Build a reliable job runner
Build one adaptable decision story
Defend your role fit
Prepare one project deep dive
Design an evaluation artifact registry
Defend a reversible launch decision
Negotiate an ambiguous role scope
Trace a project from claim to evidence
Choose between two urgent priorities
Resolve a cross-functional disagreement
Own a consequential failure
Increase a team’s leverage
Warm a cold cache without a thundering herd
Ship a 500 GB checkpoint without a startup stampede
Make every document decision traceable to evidence
Share eight GPUs between two workloads with different promises
A mutable indexed sequence that must restart exactly
Trace a falling unit price to a rising bill
Keep a very long conversation usable while it streams
Design durable scheduled publishing
Make speculative code edits safe to accept
Recover a grouping rule from two-word phrases

CONTENT BOUNDARY

Independent frontier AI interview practice. Includes adaptations of collected interview material and original exercises; not official company questions and not affiliated with the companies represented. Collected material does not predict future interview questions. Report a correction.

Loading the question catalog…

FREE PRACTICE

Public demos, always pinned.

Start with either demo. They remain here when you filter or sort.

Free account: sign in to practice the marked Free questions.

Newest published questions; deterministic, never randomized.

Demo 1Number search with delayed feedback
Demo 2Design real-time device monitoring and control

Question bank filters

Reset filters

Question groups

Filter question bank by interview groupAll 269 Coding 162 System design 48 Behavioral 29 Hiring manager 16 Project deep dive 14

Topics

Independent and not affiliated with the companies represented. Collected material does not predict future interview questions.

215 questions

Sort question bankSortFree first Newest Oldest AI ratio ↓AI ratio ↑

Question bank results

Every question has a public brief with the task, constraints, and follow-ups. Labels describe interactive practice access.

Public brief

Topic tags
Python
Data structures

xAICoding

Python codingWebAssembly + public tests

Reported interview periodFeb 2026

Weighted LRU Cache

Build an LRU cache whose capacity is the total weight of stored entries rather than the number of keys.

Public brief

Topic tags
Python
Data / SQL

xAICoding

Python codingWebAssembly + public tests

Reported interview periodJan 2026

Live Audio Room Analytics

Events have (operation, room_id, user_id, timestamp), where operation is create, join, or leave. Creating a room also places that user inside it.

Public brief

Topic tags
Python
Concurrency

xAICoding

Talk-through codingNo local grader

Reported interview periodDec 2025–Jan 2026

Sort Integers with Multiple Workers

Implement an ascending integer sort that divides sufficiently large inputs into chunks, sorts chunks concurrently, and merges the sorted runs.

Public brief

Topic tags
Python
Reliability

xAICoding

Python codingWebAssembly + public tests

Reported interview periodDec 2025

Identify Faulty Nodes with Group Tests

New public brief

Topic tags
Python

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Best Time to Buy and Sell Stock

Given daily prices, compute the maximum profit obtainable by buying once and selling once later. You may choose not to trade, in which case the profit is 0.

New public brief

Topic tags
Python
Algorithms

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Binary Search

Given a strictly increasing array of integers and a target, return the target's index or -1 when it is absent.

New public brief

Topic tags
Python

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Count Values with an Even Number of Digits

Given a list of positive integers, return how many contain an even number of decimal digits.

New public brief

Topic tags
Python
Storage & state

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Course Schedule

There are num_courses courses labeled 0 through num_courses - 1. Each prerequisite pair (course, prerequisite) states that the prerequisite must be completed first.

New public brief

Topic tags
Python

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Maximum Non-Adjacent Sum

Each value in amounts is the reward available at one location. You may select any subset, but cannot select two adjacent locations. Return the largest total reward.

Public brief

Topic tags
Python
Algorithms

ByteDanceCoding

Talk-through codingNo local grader

Reported interview periodNot available

Palindrome Integer

Determine whether a signed 32-bit integer reads identically from left to right and right to left in decimal notation.

Public brief

Topic tags
Communication

xAIBehavioral

Discussion / designResponse prompt

Reported interview periodNot available

A thin check

A story about slowing a launch because a safety check was thin.

Public brief

Topic tags
Communication

xAIProject deep dive

Discussion / designResponse prompt

Reported interview periodNot available

After launch

Where a design failed after launch, what you measured, and what you changed.

Public brief

Topic tags
Communication

xAIBehavioralAI-generated

Discussion / designResponse prompt

Reported interview periodNot available

A call you stopped

One behavioral story about stopping a decision that was moving too fast.

Public brief

Topic tags
Communication

xAIBehavioralAI-generated

Discussion / designResponse prompt

Reported interview periodNot available

A claim you refused

A culture conversation about a claim you would not publish unchecked.

Public brief

Topic tags
Storage & state
Communication

xAIBehavioralHalf AI-generated

Discussion / designResponse prompt

Reported interview periodNot available

Fast feedback and correction

Public brief

Topic tags
Python

MetaCoding

Python codingWebAssembly + public tests

Reported interview periodNot available

Maximum profit from repeated stock trades

You are given a price for each trading day and may complete as many round trips as you like, holding at most one share at a time. Report the largest total gain reachable across the whole window.

New public brief

Topic tags
Python

Scale AICodingAI-generated

Python codingWebAssembly + public tests

Reported interview periodNot available

Reconcile noisy annotator labels

Three annotators label each item and report an integer confidence. Produce the final label by weighted vote and count the items where every annotator agreed.

New public brief

Topic tags
Python

NVIDIACodingAI-generated

Python codingWebAssembly + public tests

Reported interview periodNot available

Batch inference requests under a capacity budget

A serving path receives requests in order. Pack them greedily into batches that respect both a total-size capacity and a maximum request count, then report the batch count and batch sizes.

New public brief

Topic tags
Python

CursorCodingAI-generated

Python codingWebAssembly + public tests

Reported interview periodNot available

Apply sequential line edits to a file

An editing agent proposes line-range replacements against a small document. Apply them in order, with each edit seeing the document produced by the previous one, and report the final content.

Public brief

Topic tags
Python
Algorithms
ML / AI
Product design

Coding exerciseCodingAI-generated

Practice formatRunnable coding45–70 min

Reported interview periodN/A

Build a token-budget batcher

Implement a small batching primitive, then adapt it to padded cost, bounded output, and production reasoning.

Practice focusPython, Greedy algorithms, Boundary cases, ML batching, Complexity through a runnable coding exercise.

Public brief

Topic tags
Python
Algorithms
Data structures
Product design
Communication

Coding exerciseCodingAI-generated

Practice formatRunnable coding45 min

Reported interview periodN/A

Interval calendar

Reason about half-open intervals through conflicts, room counts, merging, and free slots.

Practice focusPython, Sorting, Intervals, Heaps, Edge cases through a runnable coding exercise.

Public brief

Topic tags
Python

Coding exerciseCodingAI-generated

Practice formatRunnable coding45 min

Reported interview periodN/A

Nested dictionary transforms

Flatten and reconstruct nested data, then support path access and lists.

Practice focusPython, Recursion, Data transformation, JSON-like data through a runnable coding exercise.

Public brief

Topic tags
Python
Storage & state

Coding exerciseCodingAI-generated

Practice formatRunnable coding45 min

Reported interview periodN/A

Command-log store

Evolve a tiny state machine through undo, compaction, and snapshot replay.

Practice focusPython, State machines, Undo logs, Compaction, Persistence through a runnable coding exercise.

Public brief

Topic tags
Python
Data structures
Reliability
Data / SQL

Coding exerciseCodingAI-generated

Practice formatRunnable coding50 min

Reported interview periodN/A

Dependency-ready scheduler

Turn a dependency graph into stable ready sets, execution layers, and failure impact.

Practice focusPython, Graphs, Stable ordering, Cycle detection, Failure analysis through a runnable coding exercise.

Previous8 of 9Next

Browse all original practice

Build a token-budget batcher
Sliding-window rate limiter
Interval calendar
Nested dictionary transforms
Command-log store
Dependency-ready scheduler
Streaming metric window
Evaluation slice aggregator
Build a reliable job runner
Build one adaptable decision story
Defend your role fit
Prepare one project deep dive
Design an evaluation artifact registry
Defend a reversible launch decision
Negotiate an ambiguous role scope
Trace a project from claim to evidence
Choose between two urgent priorities
Resolve a cross-functional disagreement
Own a consequential failure
Increase a team’s leverage
Warm a cold cache without a thundering herd
Ship a 500 GB checkpoint without a startup stampede
Make every document decision traceable to evidence
Share eight GPUs between two workloads with different promises
A mutable indexed sequence that must restart exactly
Trace a falling unit price to a rising bill
Keep a very long conversation usable while it streams
Design durable scheduled publishing
Make speculative code edits safe to accept
Recover a grouping rule from two-word phrases

CONTENT BOUNDARY