Director of Data

Required

• 8+ years of data engineering, analytics engineering, or data platform experience, including leading a data function and directly managing senior engineers; experience as a company's first or primary data leader is preferred.
• Hands-on experience building or rebuilding a modern cloud data stack, with advanced SQL skills and practical experience with warehouse architecture, data modeling, transformation, orchestration, and observability.
• Experience owning data systems used by production ML or LLM products, including versioned training, evaluation, and retrieval datasets, embeddings and vector search, labeling, leakage prevention, drift monitoring, and retrieval-quality evaluation. Must be able to explain how data quality decisions affected model performance in production.
• Experience building a customer master or identity resolution model across multiple source systems, including matching logic, survivorship rules, and field-level provenance.
• Experience delivering governed business metrics used by executives and Finance, including automated reconciliation, published data-quality standards, clear remediation ownership, and the ability to explain data tradeoffs in plain language.
• Experience implementing data governance and privacy controls, including retention, deletion, lineage, role-based access, and working knowledge of CCPA and CPRA; financial transaction or multi-entity data experience is preferred.
• Within 60 days, publish one data roadmap, review it monthly with the CTO and quarterly with the founders, and assign an owner to every data initiative.
• AI assistants answer the top 20 leadership metric questions; every AI product has a team-owned evaluation set; and no model ships without testing it with AI Engineering.
• The client record covers every active client, core fields are at least 95% complete, connected-source data uses the same model, and every material field has provenance.
• Every executive KPI and Finance statement uses governed metrics; reconciliation runs weekly; unmatched payment volume falls below 1% from $9.8 million; and critical failures are detected within one hour and resolved within one business day.
• Retention and deletion are automated, deletion requests complete within 45 days with an audit record, and warehouse access is reviewed quarterly.
• Every data engineer has a level, growth plan, and on-call; no critical pipeline depends on one person; and platform spend and unit cost are reported monthly.