Harvey is the leading legal AI platform — used by Allen & Overy, PwC, and many BigLaw firms for contract analysis, document drafting, and legal research. Series E, $5B+ valuation. The interview emphasizes vertical-AI engineering, retrieval over very large legal document corpora, and the regulated-industry care of working with privileged data.
Process
Recruiter screen → 60-minute coding (Python or TypeScript) → onsite virtual: 2 coding, 1 system design (LLM-flavored), 1 craft deep-dive, 1 behavioral. Cycle: 3–5 weeks.
What they actually ask
- Design a RAG pipeline over millions of legal documents per customer
- Design a contract-analysis system that compares thousands of provisions
- Design an evaluation harness for legal-correctness (high-stakes domain)
- Coding: medium DSA, often with NLP or pipeline framing
- Behavioral: ownership, customer empathy for lawyers, regulated-industry care
Levels and comp (2026)
- SE: $185K–$255K total (cash + late-stage equity)
- Senior SE: $270K–$370K total
- Staff: $385K–$535K total
- Principal: $545K–$740K total
Prep priorities
- Be fluent in Python (ML/AI services) and TypeScript / Node.js (product surface)
- Understand RAG patterns deeply (chunking, embeddings, reranking, structured extraction)
- Brush up on legal data formats, privilege/confidentiality concerns, and audit-trail patterns
Frequently Asked Questions
Is Harvey remote-friendly?
Hubs in San Francisco (HQ) and NYC. Most engineering roles fully remote within US.
How does Harvey compare to Casetext or EvenUp?
Casetext was acquired by Thomson Reuters and serves a broader market. EvenUp focuses on personal injury / plaintiff side. Harvey is BigLaw-focused premium. Comp at Harvey is competitive at top of late-stage AI company bands.
What is the engineering culture?
Senior-heavy hires, customer-obsessed (BigLaw partners are demanding), fast-shipping. Calmer than frontier-lab pace; more product-engineering balance.