Data/Backend Engineer (Martech)
Automate your job search with Sonara.
Submit 10x as many applications with less effort than one manual application.1
Reclaim your time by letting our AI handle the grunt work of job searching.
We continuously scan millions of openings to find your top matches.

Overview
Job Description
The Mission: The "Zero-to-Insight" Engine
At Hilbert’s AI, we shorten "months" to "minutes." For our customers, the biggest friction to growth is the "onboarding tax." We want a Growth Lead to plug in their Facebook Ads or Klaviyo credentials and see Hilbert AI start generating predictive insights instantly.
We are looking for a Martech Data Engineer to build the automated intelligence layer that makes this "magic" a reality. You aren't just moving data; you are defining the universal schemas that allow our AI to understand any B2C brand's marketing engine without a single line of custom code.
The Hard Problem: Universal Semantics for a Messy World
Every company uses Martech tools differently. One brand uses "Custom Property A" for LTV, another uses a specific "Event Tag" for conversion. The hard problem is building a system that can ingest this chaos with zero custom code and near-zero configuration, yet still output standardized, high-value insights.
Our Current Hurdles:
The "Universal Schema" Challenge: You are designing the internal logic that maps fragmented 3rd-party data into a unified, AI-ready state. How do we build a model that works for a boutique bakery and a Fortune 500 retailer simultaneously?
Automated Insight Generation: We don't just want the data; we want the answer. You will be architecting the "out-of-the-box" transformations that calculate CAC, Attribution, and Churn automatically as soon as the credentials hit our system.
Agent-Ready Data: Our AI Agents are the primary users of your work. You are responsible for ensuring the data is structured, described, and indexed so that an LLM can perform complex reasoning over it without hallucinating.
Who You Are (The Profile)
The API Architect: You’ve mastered the documentation of Facebook Ads, Google Ads, TikTok, GA4, Klaviyo, Braze, or Triple Whale. You don't just know the endpoints; you know how the data actually behaves in the wild.
The "Generalist" Product Mindset: You hate "one-off" solutions. You are obsessed with building a single, configurable engine that solves the problem for 10,000 customers at once.
The Semantic Designer: You understand that for an AI, "Table 1" is useless. You have the discipline to build the metadata and schemas that give the AI the context it needs to be "smart."
Python & Airbyte Native: You’ve built custom connectors and understand the engineering rigor required to maintain "five-nines" reliability for automated ingestion.
What You’ll Own
Zero-Config Ingestion: Lead the development of our automated 3rd-party ingestion engine, focusing on "credential-to-insight" velocity.
Universal Martech Models: Design and maintain the standardized schemas for Ads, CRM, Event, and attribution data that power our core AI models.
Out-of-the-Box Analytics: Architect the logic that generates instant growth insights (Attribution, Retention, Spend Efficiency) for every new customer.
Agent Integration: Ensure all Martech data is perfectly mapped for our Agentic Query Engine, enabling natural language conversations over complex marketing datasets.
Bonus Points
Experience building "Multi-Tenant" data products or platforms.
Deep knowledge of E-commerce growth loops and marketing attribution.
Background in building or maintaining open-source Airbyte connectors.
Experience working with ML engineers to define "Feature-Ready" data.
Location
San Francisco, or Istanbul
At least 5 hours overlap with PST timezone (7am-5pm)
Compensation
Competitive salary + equity package, commensurate with experience.
Performance-based bonuses tied to project milestones and customer impact.
The Hiring Journey
Short form → Intro Call → Technical working session → Team conversations → Offer
Fast, human, no bureaucracy.
Automate your job search with Sonara.
Submit 10x as many applications with less effort than one manual application.
