Engineering log from inside Nodus.

1. 001

   How we cut V0 to four moving parts.

   The first version of Nodus is one Python workspace, four processes, two stateful stores, and roughly 1,500 lines of code that do real work. What we kept, what we deferred, and the rule we used to tell the difference.

   19 Apr 2026 · System design · 8 min

   Read →

1. 002

   The Trust Ledger schema, line by line.

   Why we hoist five columns out of JSONB, what we measure per supplier, and the index that turns the score table into a routing decision.

   Drafting
2. 003

   First real failover: anatomy of a saved job.

   From the moment a vendor returns a 500 to the moment the same job lands on a different region. What the logs look like. What we wish we had logged.

   Waiting on data
3. 004

   The boring procurement post nobody else writes.

   What it actually takes to put a Tier-3 supplier on a Fortune 500 vendor list. The forms, the calls, the months. Why this is a moat and not a chore.

   Outline
4. 005

   Why we’re not building a dashboard yet.

   The customer asks for one in every intake call. We say no, on purpose, for now. Here is the rule that decides when we say yes.

   Idea

Want to suggest a topic, or read a draft early? Tell us.