AI 服務暫時不可用，以下為來源正文，待恢復後補全翻譯。
Free to start · No card Debug your way to the offer. Real broken code. Real interview pressure. An AI coach that refuses to just hand you the answer. Practice the part of interviews that actually tests your thinking. Start practicing free →How it works ↓ 💻 The practice workspace is built for desktop. Create your account now and practice from your computer. lru_cache.py3 bugs 1class LRUCache: 2 def init(self, capacity): 3 self.cap = capacity 4 self.cache = {} 5 self.head = Node(0, 0) 6 self.tail = Node(0, 0) 7 # ← sentinel nodes never linked 8 9 def get(self, key): 10 if key not in self.cache: 11 return -1 12 # ← missing MRU promotion 13 return self.cache[key].val expected3 received-1 Scenarios styled after interviews at MetaAmazonGoogleStripeShopify Not affiliated with or endorsed by these companies How a session works Three stages. No hand-holding between them. Pick Choose a scenario Select from multiple real interview formats across Easy, Medium, and Hard. Each is a broken implementation waiting to be fixed. → Debug Find and fix the bugs A live test runner shows what's failing. An AI coach in the sidebar answers questions but won't write the fix for you. → Review Get scored on how you think Six behavioral axes, an AI-generated headline about your session, and a full timeline of every decision you made. What makes it different Not LeetCode. Not flashcards. Actual interview simulation. Debug real broken code Pre-broken implementations from actual interview patterns. Find and fix the bugs, no blank-slate busywork. AI coach, not answer machine Hints redirect your thinking toward the bug. The coach refuses to write the fix and notices when you rubber-stamp its suggestions. 6-axis performance review Scored on Diagnosis, Independence, Precision, Verification, Recovery, and Test Ownership. Not just pass/fail. Real company scenario formats EventEmitter (Meta), LRU Cache (Amazon), Prefix Trie (Google), Rate Limiter, Transaction Engine, and 6 more. Hard mode Adds realistic interviewer pressure: scope creep, over-engineering bait, correctness doubts. Practice staying focused. Live in-browser test runner Instant feedback on every edit. Hidden tests reveal edge cases once you complete each part. Real problems. 3 difficulty levels. Each one is a real broken implementation. A sample: 3 parts · 30 min EventEmitter Three bugs: listeners fire in reverse order, off() wipes the whole event, once() never removes itself. pub-subclosures 3 parts · 35 min LRU Cache Sentinel nodes never linked, eviction hits the wrong end, and get/put skip MRU promotion entirely. data-structureslinked list 3 parts · 25 min CSV Parser Drops the last field on every row, strips leading whitespace, and enters quote mode at the wrong character. parsingstrings 3 parts · 30 min Prefix Tree (Trie) search() returns True for prefixes, get_words_with_prefix() builds results in reverse, case handling is inconsistent. data-structuresrecursion + 7 more: Rate Limiter, Feature Flags, Log Query Filter, Dependency Resolver, Merge Intervals, and others. Simple pricing Practice unlimited scenarios for free. Pay only for AI coaching. Free $0/month No credit card required 2 AI-coached sessions per month Unlimited practice sessions without AI 3 practice tracks Full editor, tests, and answer keys Start free Most popular Pro $9/month Cancel anytime 60 AI-coached sessions per month All 7 tracks, 36 scenarios AI performance grading on every session Full session history and progress tracking Get Pro → Ready to debug your way to the offer? Free account. No credit card. First session in under a minute. Create free account → AI OA & Coding Interview Prep: Debug Real Code with an AI Coach