Forward Deployed Data Engineer (Integration)

Hilbert's AI•San Francisco, CA

3d•Onsite

About The Position

Hilbert is seeking a Forward Deployed Data Engineer to architect data integration solutions for B2C companies. This role involves ingesting data into Hilbert's Clickhouse managed warehouse or implementing logic directly within customer data warehouses like Snowflake or BigQuery. The primary goal is to provide AI agents with a high-fidelity view of business data, ensuring "Reasoning Consistency" across different environments. The engineer will work on a dual-track system supporting both data piping with Dagster and Airbyte, and "Zero-Extract" transformations natively in customer warehouses. A key responsibility is building the AI Semantic Layer, acting as a "Translator" to define metadata and business logic for custom datasets, enabling AI agents to understand custom columns without hallucination. The role also involves partnering with an AI Data Discovery Agent to automate data analysis, suggest mappings, and generate pipelines.

Requirements

Equally comfortable optimizing a Clickhouse query as writing native Snowpark (Snowflake) or BigQuery SQL.
Comfortable implementing data orchestration scripts using Python.
Discipline to define the "meaning" behind the data.
Ability to earn technical trust with a customer's Data Architect in minutes, extracting the logic of custom tables and mapping them to Hilbert reasoning engine.
Excited to use and improve AI agents that handle the heavy lifting of data discovery and pipeline scaffolding.

Nice To Haves

Deep experience with dbt for warehouse-native modeling.
Experience with more than one state-of-the-art data warehouse solution and knowledge of optimization strategies.
Experience with Semantic Layer frameworks (Cube, MetricQL, etc.).
Background in E-commerce/Retail (understanding Revenue Metrics, Order lifecycle, LTV, CAC, and Attribution etc.).
Having built an agentic workflow before.

Responsibilities

Own the technical lifecycle of new customers, choosing and implementing the best deployment path (Managed Clickhouse vs. Warehouse-Native).
Use Hilbert internal Discovery Agent to create reports and suggest mappings, moving from raw data to a working v1 pipeline in record time.
Architect the semantic definitions for custom enterprise data, ensuring our agentic conversation engine has the "Ground Truth" for every query.
Transform diverse source data into Hilbert unified growth models to power our generic ML systems.
Act as the lead technical resource for high-stakes enterprise implementations, ensuring our stack is "packaged" and performant in their specific infra.