Grants are not charity — they are structured R&D contracts. We approached the program the same way we approach enterprise pilots: crisp scope, measurable risk reduction, and receipts.
Hypothesis
Attackers route invalid traffic through ephemeral cloud tenants. We proposed a detector that correlates ASN churn with paid-session anomalies — something customers wanted but could not staff.
Application anatomy
- Market evidence — third-party fraud reports plus our anonymized incident stats.
- Technical plan — milestones tied to deployable services, not slideware.
- Team — operators who already carried pagers for the product.
Review day
Panelists pressed us on false positives. We showed live replay traces and rollback plans. Having production traffic already on the platform mattered more than a polished deck.
After the wire hit
We hired two infra engineers, built the streaming feature store slice we promised, and kept commercial sprints on a parallel track so the grant work did not starve customers.
