Data Architecture Agent

Our agents build audiences, run journeys, and activate campaigns in the time it used to take a marketer to file a ticket. This role focuses on setting the vision for data architecture for our most complex enterprise customers, deploying teams to build it, and ensuring quality throughout the process.

What You’ll Do

• Establish what the customer is trying to do by running discovery with marketing and data teams, translating vague goals like “we want better targeting” into specific activation use cases.
• Profile and understand the data, with agents doing the first pass, and you deciding what it actually means to architect a solution that closes data gaps at the customer level, landing them with a working Hightouch configuration.
• Create target-state schemas and identity matching rules that make a customer’s data activation-ready.
• Deeply understand how the customer targets and at what grain—e.g., person, household, device—and design identity matching rules to support it.
• Direct the build by putting agents to work profiling the warehouse, proposing models, and generating transformations. Review everything they produce, rewrite what they get wrong, and own the models and their performance regardless of who or what typed them.
• Define the number that proves it worked, agree on it with the customer before the build starts, compute it, and defend it to senior leaders.
• Hand off the day-to-day but remain the architect. You remain the named person they and their account manager come back to when the data model changes, new sources land, or their identity graph needs rework. You produce reference architectures the next engagement can reuse.
• Build the context our agents run on. The semantic layer, metric definitions, and business logic you encode let a marketer at a Fortune 500 ask a question in plain language and get an answer worth acting on. Bad context is worse than no context, and it’s yours.
• Be the eyes and ears for our product teams. You’ll see our gaps before anyone else does—bring them back sharp enough to act on.

About You

• You direct agents, you don’t just prompt them. You’ll point them at warehouses with undocumented tables and no reliable dictionary, and you’ll be the last judgment before what they build ships to millions of people. You reach for an agent before a blank editor, but you scrutinize every line before it goes out.
• SQL Fluency. Anyone can generate a thousand lines of SQL now. You’re hired because you can read them and find the join that’s wrong. Fluent in Snowflake, Databricks, or BigQuery, with query performance and warehouse cost as things you think about unprompted.
• You’re a driver. You’ve parachuted into messy engagements with dozens of stakeholders who have differing opinions and driven discovery, clarity, and alignment.
• 5+ years in a customer-facing technical role—analytics engineering, solutions architecture, forward-deployed engineering, data consulting, or in-house marketing data work at a large brand.
• This is an agent-native role. The work is data architecture. The leverage is AI. If your instinct is that doing it by hand would be faster, this will be the wrong seat.

Bonus Points

• Experience with dbt, SQLMesh, or similar transformation tooling; git-native, PR-driven workflow.
• Prior work with identity resolution, CDPs, or MDM.