Harvey Interview Question Bank
Harvey builds domain-specific generative-AI tools for law firms and legal teams, layering retrieval and document-workflow products (Vault / Data Room) on top of frontier LLMs. Its interview loop is small and stable: a single coding screen drawn from a tight rotation, system-design rounds anchored on Harvey's own document products, and a values-driven behavioral bar — with recruiters who actively prep candidates before each round.
36
questions
Last updated
Before you start
Showing 11 of 11
Implement a Read-Through and Write-Through Cache AdapterImplement an LRU Cache Using OrderedDictMini Spreadsheet with Eager Evaluation and Cycle DetectionImplement a Minimal Spreadsheet with FormulasEmbedding Retrieval with Cosine SimilarityIn-Memory Vault File System: Unique File NamesFind Duplicate Files in a FilesystemDesign an In-Memory Vault File Directory SystemHighlight Exact Source Citations in Generated TextImplement an In-Memory Spreadsheet with Addition FormulasInsert Citation Tags into TextDesign a Mini Spreadsheet with FormulasDesign an Excel-like Spreadsheet with Formulas and Dependency ValidationDesign an Excel-like Spreadsheet with Cached Formula Values and Cycle DetectionDesign an In-Memory File System with a TrieDesign a Spreadsheet with Formula Evaluation (Topological Sort)Spreadsheet with O(1) Read and Dependency ManagementDesign an In-Memory File System APIEvaluate Symbols in an Expression Map (with Cycle Detection)Live Coding: String Editing with Mouse CursorSpreadsheet with Integer and Formula Cells + Circular Dependency DetectionTag Exact-Matched Sources in a Document and Provide CitationsDesign a Versioned Key-Value Vault with Consistency Check (Filesystem-like)Citation ProblemWord Level Match and Tagging