Data Platform Engineer, Ingestion

ABOUT MINERVA

Minerva builds AI for marketing leaders. We work together with marketers to automate and accelerate their workflows with AI agents. Our agents augment the end-to-end workflow of the marketer by automating analysis, campaign planning, activation, measurement and reporting.

Customers like the NBA, Ramp, Wander, Hard Rock Stadium Group and Luxury Presence use Minerva to develop a better understanding of their customers and introduce AI specifically built to understand the customer into their commercial organizations. Our campaigns outperform other tools because we built the entire stack: from identity to enrichment to data integration as well as agents to action and report on the analysis.

We've raised $20M from The General Partnership, 8VC, Lingotto, NBA Investments, Topology Ventures, Future Positive, Background Capital and many others. Our team brings together operators and investors from Citadel, Dentsu, Bridgewater, Meta Superintelligence, Growthloop, Amazon and Lazard, alongside engineers and researchers from Berkeley, MIT, Stanford and Cambridge.

THE ROLE

Every new data source starts as someone else's problem: an API with rate limits, a vendor file that arrives late and gets corrected twice, a schema that changes without warning. Today, adding a new vendor feed still takes hand-built work.

You make new sources dependable and make the next one cheap. You build the ingestion pipelines for our customer and vendor data, automate onboarding from configuration to first delivery, and turn what you build into modules other Minerva teams can use without running platform infrastructure themselves.

This is a mid-level individual contributor seat on the data platform team, reporting to our Head of Data Platform. Our data product team decides which sources matter and what good data looks like. You own how it arrives and how it stays dependable.

WHAT YOU'LL DO

- Build pipelines for messy sources. APIs, files, database extracts and approved scraping sources, at high volume, with inconsistent formats and schemas that change.

- Make reruns safe. Incremental loads, backfills and retries that never duplicate or corrupt data. Bad records get quarantined, not loaded. Throughput and processing cost stay under control.

- Automate onboarding. The path from source configuration and access checks through provisioning, validation, deployment and first delivery runs without hand-holding.

- Build modules other teams use. Reusable connectors and storage components with tests, examples and clear interfaces, so a team building an API integration or a scraper gets managed ingestion, storage and monitoring without building them.

- Keep the origin of every value. Carry source metadata through the pipeline (which vendor, which delivery, when) so we can always say where a value came from. Enforce privacy controls so restricted data can't come back through a new load or a replay.

- Own your integrations in production. Connect each one to shared monitoring, write its recovery runbook and fix source-specific problems when they break.

OUR STACK

- Orchestration: Dagster, with Airflow still running the workloads that haven't moved.

- Transformation: dbt and SQLMesh.

- Storage: Snowflake, plus a lakehouse on Spark, Iceberg, Trino and AWS Glue.

- Code: Python on AWS.

WHAT YOU HAVE

- Typically three or more years building and supporting production data pipelines, or equivalent ownership.

- You've brought in data from many vendors and source types, including messy data at scale, and kept it running after launch.

- Practical judgment about incremental processing, backfills, failures and cost. You can show the data stayed correct as volume grew.

- You've automated repeated work or built components other engineers actually used.

- Strong Python and SQL.

- You pin down what a source owner actually needs, and you write down what you will and won't support.

- You're in New York City or willing to relocate.

NICE TO HAVES

- Connector frameworks, whether off the shelf or your own.

- Onboarding automation or self-service data tooling.

- Consumer, identity or contact data, and the privacy rules that come with it.

- Dagster, Iceberg or Snowflake in production.

YOU MIGHT BE A FIT IF YOU ARE…

- A data engineer at a data vendor or identity company who has onboarded source after source and wants to build the framework instead of the next one-off.

- An engineer at a connector, integration or reverse-ETL company who has built connectors other people depend on.

- A data engineer at a fintech, marketplace or adtech company who became the person every new integration ran through.

- Or you don't fit any of these neatly, but you've taken data you didn't control and made it dependable. Apply.

WHAT SUCCESS LOOKS LIKE

- First 90 days: a new source is live through automated onboarding you built, and another engineer has onboarded a different source with the same modules. You've measured the setup effort and tested recovery.

- First 6 months: teams outside data platform bring in new sources through your modules, and every feed you own has alerting and a recovery runbook a second engineer has used.

- The numbers you own: time to onboard a new source, and how often a feed needs hands-on recovery.

COMPENSATION

$170,000–$236,000 base salary, depending on level. Meaningful equity. Medical, dental and vision.

HOW TO APPLY

Skip the cover letter. Walk us through one pipeline you built for data you didn't control: the source, what broke, how you made reruns safe and how the next source got easier. A few paragraphs. Applications without it won't be reviewed.

WHERE WE WORK

We work in person from our office at 300 Kent Avenue in Williamsburg, Brooklyn.