SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Stand is building a next-generation property insurance platform powered by advanced physics, AI, and data-driven decision-making. Rather than pricing loss after it occurs, Stand simulates how catastrophes affect individual properties and automates underwriting and mitigation. This role owns the data foundations that power every underwriting, pricing, and operational decision across the platform.
You will architect and migrate Stand's analytics infrastructure from a centralized Postgres database (currently rebuilt hourly in full) into a modern, scalable warehouse with dbt, a real orchestrator, and per-dataset freshness SLAs. The scope includes building a trust framework and diagnostics console that automatically reconciles data across systems of record, runs tests on every load, and surfaces data health to non-technical stakeholders. You'll establish a shared semantic layer—a single registry of typed, versioned, permission-aware entities and metrics—that powers dashboards, agents, and self-service analytics. You'll also build an automated production-to-sanitized pipeline that de-identifies data while preserving distributions and joins, making realistic data available for local development and prototyping without exposing PII.
This is a foundations role measured by how well others can work independently: how fast an actuary can rate a cohort, how confidently an underwriter trusts a stage count, and how quickly an engineer can prototype without filing requests. You'll partner closely with Actuarial, Underwriting, and Applied Science teams to understand their information needs and build systems that let them solve problems themselves.
In the first 30 days, you'll justify and implement a warehouse choice, reproduce existing tables in dbt with automated parity checks, and set up orchestration with dependency graphs. Within 3 months, you'll launch a governance-first pipeline for self-serve metric definitions and a diagnostics console for non-engineers to monitor data health. By 6 months, you'll have a semantic layer enforced across dashboards and agents, a sanitized production mirror seeding all non-production environments, and leadership reporting fully automated.
