Product Manager - Data

Harmonic•New York, NY

3d•$200,000 - $250,000

About The Position

Harmonic is hiring its first Data Product Manager to own the roadmap for the data that powers its startup discovery products and AI agent, Scout. Reporting to the Head of Product, this person will work closely with engineering and data operations to decide which data gaps to address, which information sources to pursue, and how to determine whether data is good enough for customers to act on. Their decisions will shape what customers can discover in Harmonic and what its AI products can reliably do. The role involves improving data collection and acquisition, reconciling conflicting evidence into a knowledge graph, resolving company and person identities, modeling relationships and changes over time, and establishing measurable standards for accuracy, coverage, and freshness. The Product Manager will lead new data capabilities from initial scoping through adoption, investigate data problems, and prioritize systemic fixes. Harmonic combines internet-scale collection, information shared directly by founders, partner data, and internally derived insights. Harmonic describes itself as a startup discovery engine used by investors, corporate development, and go-to-market teams. The company says it is growing revenue at more than 100% annually and has received $30 million in funding from Craft, Floodgate, and Sozo Ventures.

Requirements

At least 5 years of product management experience, including recent responsibility for data products.
Experience owning a data product and explaining how product decisions improved customer capabilities as well as underlying data metrics.
Experience working with company, people, or financial data.
Ability to investigate conflicting information about a company and reason about how to address the problem across large-scale datasets.
Technical fluency with data ingestion, extraction, entity resolution, and classification, and ability to work through these topics with engineers.
Ability to independently investigate data to test assumptions.
Ability to make informed tradeoffs among accuracy, coverage, freshness, latency, and cost.
Experience building or owning evaluation frameworks and understanding their limitations.
Ability to select representative samples, establish ground truth, and determine whether an improvement is ready to ship.
Ability to make clear decisions under uncertainty, focus a roadmap, explain priorities, and define useful initial releases.

Nice To Haves

Experience at a private-market data provider or building data-intensive workflows for investors.
Early-stage B2B experience.

Responsibilities

Identify data improvements that would most change what customers can find, evaluate, or act on, using customer research, usage, commercial opportunities, and quality issues to set investment priorities.
Decide whether to collect, partner for, request from users, or derive information, evaluating sources for incremental value, reliability, cost, and freshness.
Partner with the founder-facing team to make direct data contributions useful to founders and customers.
Work with engineering to reconcile conflicting evidence, resolve company and person identities, and model relationships and changes over time.
Determine how uncertainty and data provenance should affect what customers and Scout can rely on.
Define and maintain evaluations for accuracy, coverage, and freshness that reflect customer tasks; establish baselines and launch criteria and catch regressions.
Assess whether aggregate quality scores obscure important gaps in customer-relevant segments.
Lead new data capabilities from idea through adoption, potentially including mapping company–investor relationships over time, adding cap table and revenue data, and building AI agents that extract structured information from difficult sources.
Scope useful initial releases with engineering and work with product and go-to-market teams to deliver them to customers.
Investigate data issues raised by customers, internal teams, and auditing agents; identify whether they stem from collection, extraction, matching, or modeling; and prioritize fixes that improve the dataset beyond a single record.
Stay close to customer workflows and disputed records, and follow through to determine whether shipped work solved the problem.