2 min lesson
Generalize the one case into a primitive
Give a practical answer to this: "In the live design, why does a waterfall enrichment usually beat hitting all providers in parallel and what one number must you cap regardless of which approach you choose?"
Step 1 of 3
Generalize the one case into a primitive
The last five minutes are where this round is actually won. The prompt was one program, but Cursor wants reusable program primitives that extend across segments, geos and program types.
Enrich → score → route for ONE startup program.
Hardcoded thresholds and one routing map.
A scoring primitive parameterized by signal weights + thresholds.
A routing primitive parameterized by ownership model + SLA, so partnerships and dev-ecosystem reuse it with config, not a rewrite.
In AI-barred mode, narrate at least two options you rejected and why - "I considered a fan-out to maximize match rate but rejected it on cost; I considered an ML score but rejected it until we have labeled outcomes." Stating the road not taken is the clearest signal of senior judgment and it's the thing silent diagrammers never do.