Data & Analytics Engineering in NYC & Northern NJ
We build the data infrastructure that lets a business answer its own questions. Warehouses, pipelines, and dashboards that pull from your operational systems, accounting, ad platforms, and the spreadsheet that's still load-bearing — and present a single set of numbers your accountant, your banker, and your operations lead will all agree on.
What we actually build.
Data Warehouses
Snowflake, BigQuery, or Postgres warehouses sized to your scale. Schema designed for the questions you actually ask — operating reports, board materials, cohort analysis — not a generic star schema lifted from a tutorial.
ETL & Reverse ETL Pipelines
Pulls from QuickBooks, NetSuite, Salesforce, HubSpot, Stripe, ad platforms, and your application database. Reliable schedules, monitored failures, and source-truth documentation. Reverse-ETL back to operational tools where it matters.
Operating & Financial Dashboards
The reports your leadership team actually uses every week — built once, maintained quarterly, and trusted because the numbers tie out to your books and operations data.
Self-Serve Analytics Layer
Semantic models in Cube, dbt Semantic Layer, or Looker so your team can answer their own ad-hoc questions without re-deriving the same metrics inconsistently across tools.
The process.
- 01
Inventory
List every system that holds operational or financial data. Document who owns each, how often it changes, and which definitions of common metrics differ across them.
- 02
Model
Define the business entities (orders, customers, accounts, employees, transactions) once, with written definitions. The model is what stops the 'whose number is right?' arguments.
- 03
Build
Pipelines into the warehouse, transformations in dbt, dashboards in your BI tool of choice. Tests on the data so you find out about a broken pipeline before your CFO does.
- 04
Operate
Daily reliability monitoring, quarterly schema refresh as the business evolves. Documentation that an in-house analyst could pick up if you hire one.
NYC & Northern NJ in person.
Data engagements work well partially onsite. The kickoff sessions where you walk us through how the business actually flows are dramatically more productive in person — most operating businesses have meaningful institutional knowledge that nobody has written down. We're headquartered in NYC and run kickoffs across the five boroughs same-week. For Northern New Jersey clients we schedule those sessions on the days we're in-state (twice a week), including time at the offices of the operations and finance leads. The pipeline and warehouse work is then largely remote, with periodic visits for major reviews.
The stack.
- Snowflake
- Postgres
- BigQuery
- dbt
- Dagster / Airflow
- Fivetran / Airbyte
- Cube / Looker
- Metabase / Hex
Warehouse on Snowflake (mid-market), Postgres (smaller scale, simpler), or BigQuery (Google-shop). Transformations in dbt. Orchestration in Dagster or Airflow depending on team preference. Ingestion via Fivetran or Airbyte where it makes sense, custom connectors where it doesn't. BI in Metabase, Hex, or Looker.
Who we work with.
Multi-system operators
If you have separate systems for accounting, CRM, e-commerce, ads, and operations — and reconciling them is somebody's monthly nightmare — a warehouse pays for itself.
Companies preparing for a fundraise or sale
Diligence-grade financial and operational reporting. We've built the data rooms that get a deal across the line.
SaaS companies past PMF
Real product analytics, cohort and retention reporting, channel attribution. The metrics your investors and your engineering team should be able to look at the same way.
Common questions.
- Do we really need a data warehouse?
- If you have one operational system and a small business, probably not. If you're reconciling data across three or more systems weekly, yes — and you've been working around the absence of a warehouse for longer than you should have.
- Snowflake or Postgres?
- Postgres until you have a real reason to leave (cost of long-running queries, scale, separation of concerns). Snowflake when you do. We'll write down which is right for you and why — it's one of the most over-prescribed decisions in the industry.
- Can you make our existing dashboards faster?
- Often, yes. Most slow dashboards are slow because they're querying operational systems directly or recomputing the same metric in five places. Moving the heavy lifting into a warehouse and a semantic layer fixes both.
- Will my accountant trust these numbers?
- That's the requirement. We tie financial metrics back to the source-of-truth system (QuickBooks, NetSuite, etc.) and document the reconciliation. If the numbers don't tie, the dashboards aren't done.
- Can you train my team to maintain this?
- Yes. We can deliver as a turnkey product, as a co-built engagement with your team, or as a training engagement. Documentation and runbooks are part of the deliverable in all three cases.
The other eight.
Start a conversation.
Direct reply from the founder. NYC & Northern NJ in person; U.S. clients remotely.
Get in touch →