Line 03 / Research
BlackGrid Labs builds deterministic security tooling — fraud detection, identity verification, tenant isolation — that organizations can audit, reproduce, and trust. We work where the conventional approach is opaque, probabilistic, or hostage to a vendor.
Inbound mail, payment artifacts, and identity claims scored by reproducible signals — not by a probabilistic classifier deciding the verdict. Same inputs always yield the same score, every decision is auditable.
Email threat intelligence →128-dim biometric descriptors encrypted at rest, Euclidean-distance matching server-side only, gaze-rotation liveness signatures that distinguish real users from photo, video, or generated faces. No vendor lock-in.
Identity & liveness →A single bug in application code should not leak data across tenants. The guarantee should hold at the database layer, not the application layer. Schema-level row policies + an adversarial fuzz harness, 4M+ queries, zero leaks.
Isolation services →Run 014 of liveness-rotation yielded 99.2% reject rate on synthetic-face dataset v2. Pushing the threshold to the MyMeridi MFA pilot in week 21.
Track 02D.A.E. email-auth crossed 91.4% on the held-out validation set. Triaging false positives.
Track 01Tenant-isolation fuzz harness completed 4.0M queries across five attack categories. Zero cross-tenant returns. Schema coverage now at 100%.
Track 04Operator anomaly: opened the dataset for the first batch of consenting tenants. Rule lexicon at 12 deterministic rules; targeting 30 by end of phase 1.
Track 03Talk to the lab
If your operating problem looks like one of our research tracks, we want to hear about it. The lab works most productively with operators who can describe their friction in technical terms.