10 days with Claude Code → an AI analyst that physically can't make a number up

I built a production-ready AI analyst in 10 days with Claude Code — but it really took 5 months and a second brain to get there. 🍑

The honest version of this story: Plum didn't come out of nowhere. It was the payoff of a Feb–June learning sprint — SQL, Python, Power BI, Tableau, Databricks, then Claude Code and AI agents — where I built a medallion warehouse three times over in different stacks before I ever pointed it at a real product.

What tied it together was a second brain in Obsidian. Not a pile of notes — a structured wiki that Claude Code itself maintains: one page per course and project, skill pages that trace every claim back to the work that proves it, an append-only log, and a hard rule that contradictions get flagged, never silently overwritten. So when I started Plum, I wasn't starting cold — I was starting from a connected map of everything I'd actually built and could reuse.

Then I pointed Claude Code at a hard problem: an AI analyst for Shopify whose entire design goal is restraint — it physically can't make a number up.

The build (Claude Code, ~10 days):

→ A real bronze → silver → gold warehouse on PostgreSQL, built from store data

→ A governed metric layer where every number traces to a defined SQL query

→ A read-only tool layer the model can only propose calls against — it never writes SQL, never touches the database

→ Receipts on every answer: the exact SQL plus a hash you can replay without the AI in the loop. Doctor a receipt, it fails.

→ A numeric post-check on the model's prose — two strikes and it's withheld, facts stand alone

→ Built by a chartered multi-agent team, verified gate by gate, multi-tenant, Shopify-App-Store-ready

The thread running through both the vault and the product is one rule: don't fake "done," don't fake numbers. A task is complete only when the artifact exists, runs, and the output was actually observed. Every metric traces to a real query. That discipline is the only reason a 10-day project is honest enough to show publicly.

Real verbatim output (not a mockup):

RECEIPTED FACTS

(1) total_revenue (all time) <rcp_c60c... | sha256:371834e1...>

total_revenue = 29,356,250

replay: python plum_replay.py rcp_c60c...

postcheck: PASS

Built in public, pre-launch, every claim reproducible: https://gooliverani.github.io/plum-site/

A few things I'd love this group's take on:

Anyone else running a second brain that an AI agent maintains for you? Mine has Claude Code write and cross-link every page under a strict schema — curious how others structure the vault ↔ build loop.

How do you keep Claude Code honest on long projects — charters, hooks, tests? Mine leaned on dev-time hooks, a CI guard, and an audit log of every AI action.https://gooliverani.github.io/plum-site/

2 comments

10 days with Claude Code → an AI analyst that physically can't make a number up