About the role
Hilbert is a scalable, data science-first growth engine that gives B2C teams predictive clarity into user behavior, revenue drivers, and the actions that drive sustainable growth. Fully agentic by design, Hilbert shrinks months-long decision cycles to minutes. From Fortune 10 enterprises to beloved brands like FreshDirect, Blank Street, and Levain Bakery, operators run their growth on Hilbert. The Mission: Architect the Ground Truth At Hilbert we shorten "months" to "minutes." We build the AI Growth Engine that Fortune 500 enterprises and iconic B2C brands use to navigate their revenue drivers in real-time. We are expanding our deployment model. Whether we are ingesting data into our high-performance Clickhouse managed warehouse or implementing Warehouse-Native logic directly within a customer’s own data warehouse solution like Snowflake or BigQuery, the goal is the same: providing our AI Agents with a perfect, high-fidelity view of the business. We need a Forward Deployed Data Engineer who can architect these bridges in any environment. The Hard Problem: Architectural Hybridity & Semantic Clarity The challenge isn't just moving rows; it’s ensuring "Reasoning Consistency." You are responsible for making sure our AI produces the same high-quality insights regardless of where the data resides. Our Current Hurdles: The Dual-Track Challenge: We are building a system that can either pipe data into our Clickhouse stack using Dagster and Airbyte , or execute "Zero-Extract" transformations natively in the customer's warehouse. You’ll need to master both worlds. The AI Semantic Layer: For customers with custom datasets that don't fit our canonical models, you will build the "Translator." You’ll define the metadata and business logic so our agentic flows can "understand" custom columns without hallucination. Agent-Assisted Integration: We are building a Data Discovery Agent to automate the "messy" part of your job. You will prompt and partner with this agent to analyze customer data, suggest mapping alternatives, and generate the first version of pipelines automatically, slashing integration time significantly.