Questions an interviewer could realistically ask about this project, drawn from
../learning/. Sorepoint scans local businesses, finds each one’s single worst
fixable flaw via six competing “agents”, and builds + mails the fix.
How to use this
- Answer out loud, from memory, before reading the hint.
- Strong answer covers = the checklist a good answer hits, not a script.
- [Beginner] / [Intermediate] / [Advanced] = how much project context the question assumes.
- 🔥 = most likely to be asked.
Files
| File | Covers |
|---|---|
| 01-pipeline-and-crawling.md | Worker-not-route, source adapters, OSM, cache-first/resumable, the crawler, honest failures |
| 02-postgres-and-supabase.md | Local stack on rootless Podman, migrations, TypeGen, GRANTs vs RLS, idempotent upserts |
| 03-product-decisions-and-story.md | Honest UI, evidence-weighted ranking, the ethics of unsolicited outreach, behavioural |
Before anything else: the 60-second pitch
Sorepoint scans local businesses in a city+niche, and for each one identifies the single worst
fixable flaw — six agents compete on evidence, one wins — then builds the fix and mails it. The
pipeline runs as a standalone worker script writing to Postgres, not as an API route, because the
external calls outlive a serverless request and have to be resumable. Two things I’d defend
hardest: every source sits behind one normalised interface, which is what let us swap Google
Places for OpenStreetMap in an afternoon when billing blocked us; and the whole system is built
to be honest about what it doesn’t know — a signal we couldn’t determine isnull, never a
defaultedfalse, and a site that blocked our crawler is recorded as evidence about us, not as
a flaw of theirs.
The “null, not false” line is the one to lead with. It’s a one-sentence demonstration of data
integrity thinking.
Software engineering
Infrastructure & security
Computer science
Hardware & engineering