Loading...
SeniorFull-timeAI
Salary Range
$143,650 - $194,350 /year
This salary is estimated based on similar roles. The actual salary may vary.
About the Role
We are looking for a Product Manager to join Harvey, a company transforming how legal and professional services operate through frontier agentic AI and an enterprise-grade platform. This role focuses on leading the next generation of Harvey’s data engine as we scale our capabilities.
Key Responsibilities
- →Drive the roadmap and strategy for Harvey’s ‘Data Factory’, ensuring we scale our data 100x through new platforms that build the ‘legal index’ of the world
- →Work with internal operations and external data providers to expand coverage, accelerate execution, and improve dataset quality
- →Own and evolve Harvey’s end-to-end data architecture, ensuring performance, reliability, and scalability for LLM-powered products
- →Partner with Applied AI engineers to build and optimize retrieval systems, embeddings, search models, and evaluation frameworks
- →Architect and oversee large-scale ingestion pipelines that aggregate, normalize, and continuously update millions of heterogeneous legal documents
- →Collaborate cross-functionally with Product Engineering, Applied AI, Research, and Platform teams to deliver high-quality production systems
Requirements
- ✓5+ years of experience building or managing search, retrieval, recommendation, or data platforms at scale
- ✓Experience working with complex, heterogeneous, or domain-specific datasets with structured + unstructured data
- ✓Understanding of modern retrieval methods, including hybrid search, dense retrieval, re-ranking, embeddings, chunking strategies, and index optimization
- ✓Hands-on experience with LLMs or RAG frameworks (evaluation, grounding, hybrid pipelines, query rewriting, LLM-as-a-judge, retrieval metrics)
- ✓Ability to partner with engineers on technical architecture, with enough depth to challenge assumptions, propose solutions, and influence design
- ✓A product mindset for search—balancing user needs, domain complexity, and system constraints to propose high-leverage improvements
