Senior Data & AI Governance Engineer

Position Summary:

Loansure serves the US mortgage industry. We license property and consumer data, receive loan outcome data from lender clients, and use both to build audiences, scoring models, and AI agents that help lenders fund more loans. Data is the product, so how it is governed directly affects what we can build and sell.
The Senior Data & AI Governance Engineer makes that data safe to use and easy to trust: who can reach it, what we are permitted to do with it, whether it is documented, and how a change in one place affects everything downstream. A working foundation is in place in Snowflake. This role turns it into a program that holds up under client security reviews, emerging AI regulation, and third-party due diligence. This is a hands-on, senior individual contributor role reporting to the CTO, working across engineering, data, product, and performance marketing.

Position Responsibilities:

• Design Snowflake role-based access control for people, applications, and AI agents, and manage roles, grants, and policies as code with peer review rather than manual changes.
• Maintain semantic views as the single governed read path for analytics and AI; certify semantic models and business definitions written by the data and BI teams.
• Run quarterly access reviews with named approvers and retained evidence, and enforce separation of duties between who builds, who approves, and who deploys.
• Detect and close access gaps created by ingestion and tooling, such as grants lost when tables are recreated.

• Build and maintain the data inventory and data rights register: for each source, its origin, contract, permitted and prohibited uses, sensitivity, retention, and approved consumers.
• Enforce purpose limits in the data architecture, such as keeping client-supplied data separate from licensed data where the terms of use differ.

• Run the production review gate for AI agents, MCP servers, and skills: what data each may reach, through which role, and with what evidence.
• Maintain the data side of the AI system inventory: which systems use which data, and whether they influence decisions about consumers.
• Trace training data and model features back to permitted sources, in partnership with the data science team.
• Set the data standard that citizen-built apps and dashboards must meet, tiered by the impact of their output, and verify apps against it.

• Maintain the business glossary, data dictionary, and lineage in a form both people and AI tools can use.
• Analyze source and schema changes before they reach dashboards, models, or agents, and document the impact for affected teams.
• Define and report data governance KPIs.
• Document every control as a runbook another engineer can operate, so no control depends on one person.
• Maintain a governance evidence package that answers client security reviews and due diligence requests.

• 30 days: current access model mapped from Snowflake roles, grants, and access history; existing controls, open governance tickets, and past decisions documented; priority data domains and their stakeholders identified.
• 60 days: runbooks written for every existing control; access-as-code tooling selected and the approach approved; agent review gate defined and agreed with the agent team; data inventory started for the priority domains.
• 90 days: one schema or role family managed as code with peer review; agent review gate applied to new agent and MCP releases; data inventory and rights register drafted for the priority domains, with open items noted where contracts or counsel input are pending; first quarterly access review completed.
• 12 months: all roles and grants managed as code; sensitive columns classified and enforced with tags and masking; quarterly access reviews running with evidence; glossary and lineage covering the semantic layer; AI system inventory current.

• 6+ years in data engineering, data platform engineering, or analytics engineering, including hands-on work implementing access controls, masking, and data classification.
• Expert SQL and Snowflake: role hierarchy design, secure and semantic views, object tagging, masking and row access policies, access history, and object dependencies.
• Experience managing access as code (Terraform, schemachange, or similar) with CI-based review.
• Python for automation and analysis.
• Clear writer who can turn policy into enforceable controls and explain them to non-technical data owners.
• Plus: governing data used by AI or ML systems, such as semantic layers, text-to-SQL, LLM agents, MCP, or feature pipelines.
• Plus: built a governance program from an early stage at a small or mid-size company, without a dedicated platform or team.
• Plus: regulated consumer or financial data (GLBA, FCRA, state privacy laws) and third-party data licensing terms.
• Plus: Snowflake Cortex Analyst, or moving semantic definitions from a BI tool into a warehouse-native semantic layer.
• Plus: NIST AI Risk Management Framework or DAMA-DMBOK; mortgage or lending industry experience.