Book a call

Data you can count on

Pipelines, reporting and back-office automation for funded startups and growing companies.

Book a call Send a message

All your data in one placeSynced on your schedule

Every number is tested firstOne definition per metric, checked nightly

Dashboards and questions, one sourceSo the answers agree

The routine work runs itselfRules for the predictable, AI for the rest, a person for what matters

What we build

Every company has jobs that get done only because one person remembers. Copying, checking, chasing, filing.

Salve turns each job into an automation you can watch run. Rules handle the routine, AI reads what rules can't, and a person decides anything that matters. Every step is logged.

The board deck says $4.1M. The CRM says $3.8M. Finance has a third number. The meeting becomes an argument.

Every source lands in one warehouse. There it's cleaned, tested and modeled, with one definition per number. Dashboards and AI tools read from the same place, so the answers agree.

Not a new idea. That's the point. This is dimensional modeling. Ralph Kimball laid it out in The Data Warehouse Toolkit in 1996, and data teams have used it ever since.

The tools changed. The method didn't, because it works. Every number traces to its source, every definition is written once, and tests catch problems before your board does.

Standard tools, so a future hire can take over

SnowflakedbtDagsterPostgresPower BI

Who it's for

Funded startups

Growing faster than their spreadsheets, with no one whose job is the data.

Roll-up groups

New acquisitions on different systems that need to report as one company.

Modernizing businesses

Established companies evolving manual processes into systems that run on their own.

How we work

EmbeddedMost common

A set number of hours each month, in your Slack and your tools, with the same person every time. A pipeline one week, a broken report the next. Change the hours or stop with a month's notice.

  • Billed monthly. Hours tracked and shared.
  • Same-day replies. Fixes before you notice.
  • Starts at 20 hours a month.

Per project

One defined problem, a written scope, a fixed price. Then a short retainer to keep it running, or a documented hand-off if you'd rather own it.

  • Scope and price before any money moves.
  • Weekly updates. A working version you can try early.
  • Typical build: four to eight weeks.

Your engineer

Justin Dickey

Justin Dickey

Data and automation engineer

Justin builds the data systems operations teams depend on. That means integrations with Salesforce, NetSuite, Workday Prism and clinical record systems, document pipelines that read invoices and route email, and the Temporal and Dagster infrastructure that keeps it all running. You talk to him directly, from the first call to the last fix.

LinkedIn

Tell Salve what takes too long or doesn't add up.

A few lines is plenty: the job, the systems, who does it now. You'll get a reply within one business day. Or skip ahead and book a 30-minute call.

Book a call