Company
About groundre
We build the independent layer that checks what AI says about a binary — so security teams can use AI without inheriting its confident mistakes.
01Why we exist
AI reverse-engineering tools write fluent, confident reports — and quietly invent addresses, flag dead code as malicious, and cite IOCs that never appear in the sample. Acting on a hallucinated finding burns hours and erodes trust in the tooling.
groundre is the deterministic check. We run Ghidra and radare2, extract ground truth once, and grade every claim against it. The core rule is non-negotiable: no LLM ever decides a verdict or a score.
02How we think about it
- The model is graded, never the grader.
- Every finding cites a concrete address, xref, section, or observed syscall.
- Unprovable is UNVERIFIED — never a false flag.
03Team
We’re a small team of reverse engineers and platform builders. Credit for the work lives in the code and its commit history — not on a marketing page.