Have a tricky problem that you think requires a rewrite of your software? Join as a guest on our podcast and we’ll talk you out of it.

In their words

“We first began working with YGG in 2025 to document the architecture behind our internal system. Aside from the thorough documentation, YGG also came to us with a number of suggestions to enhance the operation of our system. Since early 2026, they have been working alongside our in-house developer to streamline system functions. Now, jobs that would take up to five minutes take merely seconds.”
Bryce C. Harris, Universal Gem

How it went

First month: a production slowdown, diagnosed from the data

The portal slowed every time critical vault processes ran. Rocco restored the database locally, traced the job, and found a connection storm: one process and one database connection per location, about 460 at once, every run. The fix was a concurrency cap of eight at a time plus timing logs, which their IT lead shipped the same week. The job now finishes in about 30 seconds and the load spikes are gone from Datadog. The read replica in the statement of work was never needed.

Months two to four: a finance audit tool from a one-page brief

Finance gets a paper ticket for every cash can pulled from a machine and needed to reconcile them against vault counts and machine meters. Rocco built the entry and audit screen with unit tests on the calculations, then iterated weekly on what finance saw: cash in versus voucher in, prior meter readings, machine swaps, historical machine numbers, a variance-result field, per-vault tabs, Excel exports, and a paginated summary that replaced a page taking over three minutes to load.

Months four to ten: the sales CRM, rebuilt for the road

Search that no longer fired a query on every keystroke, a mobile layout the field team actually uses, a prospects stage, bulk notes, calls and tasks, admin-only controls, and other user-friendly tools which have proven impactful in the field. It replaced a paid export-and-import mapping workflow.

What they got

  • A vault job that runs in about 30 seconds instead of degrading the portal, with no new infrastructure
  • A revenue audit workflow for finance, with tested calculations, exception views and Excel exports
  • A sales CRM the sales team uses on their phones; management reported engagement up and ideas coming in
  • Quarterly commission statements in the portal and partner portal
  • A Dockerized local environment, a database-change log and a pull-request habit on a codebase that had none
  • A self-service analytics plan on Metabase so leadership can build their own reports

How the engagement runs

One YGG engineer, one weekly sync with their IT lead, and a ticket queue written by finance, sales and operations. Every change goes through a pull request in their repository and ships in their release window. Database changes are written down before they are run. When a ticket’s estimate is wrong, the client hears about it the same day. When a ticket shouldn’t be built, the engineer says so and why.

That is what forward-deployed looks like on a legacy Rails application: the work happens inside the client’s constraints, the people who use the software set the priorities, and the client’s team owns every line when it lands.

Have an internal platform like this?

Book a call. We’ll talk through what an embedded engineer would pick up first.

Book a call How we work & pricing

Prefer email? start@ygg.llc