Director, Data Product Engineering

Natera is seeking a product engineering leader to build and lead the team that designs, delivers, and operates domain data products and AI-enabled analytical solutions on NDP (Natera Data Platform). You will report to the Head of Data & AI and partner with platform, governance, and product functions to turn data needs into certified, production-grade data and analytics products.

Natera follows a data mesh architecture with headless data products: domain-owned assets, not tied to any BI layer, built for both human and AI consumption. A data product is ready when it is semantically correct and AI-ready, not just numerically accurate. Your mandate is to make that standard repeatable across every domain while transforming how the team works: AI-native, 3–5x more productive, and self-service for the business.

Please note that this is role focuses on product side of data engineering (not platform). This person needs to demonstrate their ability to

(a) Build a self-service analytics product experience for business users and

(b) Create a catalog of AI ready gold-standard cross functional data products

(c) Create an operating model focusing on reusability, speed. and business value of data

that scale beyond one business domain.

Critical Priorities for This Role

• Standardize and scale how data products are built — one publishing standard, golden paths, and a common operating model across every domain.
• Make the team AI-native — agentic SDLC and AI-assisted development become how everyone works resulting in higher productivity and growth opportunities
• Deliver a 3–5x productivity gain — instrumented with baselines and delivery KPIs, not anecdotes.
• Raise data product quality — semantically correct, AI-ready, observable, ship-ready data products.
• Make analytics self-service — certified, discoverable products that business users and AI systems consume without an engineering queue.

1. Own Data Product Delivery

• Lead the build of cross functional data products, analytics experiences, and Golden KPI's that are essential to making data driven decisions across the business
• Create the operating model to deliver analytics to business users by leveraging embedded data/AI engineers with each business domain.
• Track and report delivery KPIs: time-to-delivery, certified dataset count, adoption by consuming teams, open production incidents.

2. Standardize & Scale How Data Products Are Built for Analytics and AI use

• Define and enforce the publishing standard for analytics products: semantically correct, AI-ready, lineage documented, ownership assigned, catalog entry complete.
• Establish the headless data product standard: domain-owned assets any authorized consumer — dashboard, workflow, or AI system — can use.
• Build golden paths (templates, examples, documented patterns) so every product starts from a known-good baseline, and make this the operating model for intake, build, certification, and support.

3. Build an AI-Native Engineering Competency

• Drive agentic data engineering as the default: agentic SDLC, AI pipeline generation, agents writing and validating dbt models, AI-driven testing, LLM tools in code review and documentation.
• Train every engineer to work AI-natively and redefine the working model — SDLC steps, roles, human-in-the-loop checkpoints, definition of done.
• Own a measurable plan to lift productivity 3–5x: baseline throughput and cycle time, instrument each change, report outcomes.

4. Raise the Data Product Quality Bar

• Enforce engineering standards on every production solution: CI/CD for pipelines, infrastructure as code, data observability, data quality frameworks.
• Implement agentic data quality: automated drift detection, root-cause identification, human-in-the-loop recovery.
• Meet HIPAA, RAQA, and data classification requirements at design time. If a product does not meet the bar, it does not ship.

5. Make Data Self-Service for Humans and AI

• Own the NDP contribution and discovery model: what gets published, how it is documented, and how human and AI consumers find and evaluate certified assets.
• Ensure every catalog entry carries semantic metadata and AI-readiness classification so business users and AI agents answer their own questions without an engineering ticket.
• Partner with the Data Governance Lead to embed data contracts, certification review, and access policy into the lifecycle so self-service is safe and correctly scoped.

6. Lead the Team and Set the Technical Bar

• Hire, coach, and grow data and analytics engineers; set clear expectations, give direct feedback, build real career paths.
• Stay hands-on: review architecture and code, debug production failures, and set the standard by example across Snowflake, AWS, Claude, Sigma, dbt, Fivetran, Python, and Airflow.
• Be the engineering face of data products to business stakeholders: translate ambiguous needs into scoped, time-bound commitments; when things change, communicate early with a plan.

Required

• 10+ years in data engineering, 5+ leading data or analytics engineering teams at Director level.
• Hands-on depth: you write and review Python and SQL, critique dbt models, debug pipeline failures, and make architecture calls yourself.
• Shipped data products to production with measurable adoption. You can name the products, who used them, and what changed.
• Regulated-environment delivery (healthcare, life sciences, diagnostics, pharma) with PHI and real HIPAA compliance experience.
• Modern data stack: Snowflake, AWS, Claude, dbt, Fivetran, Sigma, orchestrator such as Airflow or Dagster; CI/CD for data pipelines and infrastructure as code.
• Working knowledge of data mesh and headless, domain-owned data products built for human and AI consumers.
• Built or led teams using AI-assisted development in production: agents, code generation, AI-driven testing and validation.
• Defined engineering standards, golden paths, or operating models that scaled across multiple teams or domains.
• Strong communicator: runs a stakeholder review, writes a technical proposal, represents engineering with executives.

Nice to Have

• Diagnostics, genomics, or clinical data engineering (BAM, VCF, FASTQ).
• Vector databases, embeddings, or RAG architectures in a data engineering context.
• Data contracts as engineering artifacts: schema enforcement, versioning, change management.
• Rolled out enterprise agentic tooling (e.g., Claude Code) to an engineering org, including training and change management.