(943) 243-0151 · contactus@axilayerai.com
axilayerai.com/govprofile
Independent AI Assurance for Government
Company Overview
AxiLayer AI, Inc. is a Delaware corporation and an independent AI assessment body. We audit, validate, and continuously monitor artificial intelligence systems for government agencies, defense components, and the prime contractors that serve them. Registered in SAM.gov for all award types since 3 June 2026 under CAGE Code 20JV1 and UEI CB76ENDLMUC9.
Our defining constraint is structural: AxiLayer AI holds no reseller agreements, no implementation partnerships, and no revenue relationship with any AI vendor, model provider, or cloud platform. We do not build AI systems and then assess them. Assessment is the entire business — which is what makes our findings defensible in front of an Inspector General, GAO, a court, or a congressional committee.
Practice coverage spans the full assurance lifecycle across eight service lines, six continents, and forty-seven-plus regulatory frameworks, delivered by certified human auditors and instrumented by AxiSentinel™, our patent-pending continuous compliance platform.
Delivery Model — Human Auditors + AxiSentinel™ Platform
Every engagement combines both layers. AxiSentinel watches continuously and signs nothing: AXI-Node agents intercept every AI inference event, evaluate it against encoded regulatory logic in under 20 ms, and raise a Provisional Alert. A certified AxiLayer AI auditor then reviews the underlying evidence, validates or overrides the alert, and applies a cryptographic signature. Only then is a Compliance Passport issued into the tamper-evident evidence chain. No certificate can be issued by automation alone — the constraint is architectural, so an agency can always name the qualified person accountable for a conclusion.
Core Competencies
- Independent AI System Auditing — third-party examination of data lineage, model behavior, decision logic, control environment, and human oversight, with an evidence file built for IG, GAO, and litigation review.
- Algorithm Assurance & Model Validation — performance, robustness, bias and disparate-impact testing, including frontier, multi-agent, and tool-using architectures.
- AI Risk Assessment — risk identification and prioritization mapped to NIST AI RMF 1.0 functions and ISO/IEC 23894.
- Compliance Readiness & Gap Analysis — posture assessment against binding obligations with a prioritized, costed remediation roadmap.
- AI Validation & Verification (IV&V) — independent verification reported directly to the government, separate from the developing contractor.
- Continuous Compliance Monitoring — post-deployment surveillance via AxiSentinel, including drift, guardrail-breach, and human-oversight-gap detection.
- Regulatory & Policy Advisory — AI requirements language for solicitations, SOWs, and vendor evaluation criteria.
- Documentation & Evidence Packages — technical files, model cards, impact assessments, oversight logs, Compliance Passports.
AxiSentinel™ Platform — Government Capability
- Air-gap and classified support. The .axibatch binary format produces continuous audit evidence inside disconnected facilities with no outbound connectivity — a capability most commercial AI governance tooling does not have.
- Cryptographic evidence chain. SHA3-256 Merkle-linked, tamper-evident, with post-quantum cryptographic agility and a 26-field Audit Telemetry Event schema.
- Human Oversight Gap Score (HOGS). Continuous five-dimension measurement of where human control over AI decisions is thinning.
- Live Certification Status API. OCSP-equivalent certificate status — VALID, CONDITIONAL, SUSPENDED, REVOKED — resolved in under 100 ms.
- Active AI threat defense. Adversarial input detection, model-update poisoning prevention, drift and guardrail monitoring, AI phishing and synthetic-media detection.
- Deploys where the mission runs. Kubernetes and cloud sidecar, serverless extension, on-premises binary, mobile SDK, Open RAN network function, and air-gapped batch.
- Developed at private expense. Three USPTO provisional applications, 37 claims, priority 6 June 2026 — all filed before any government funding or SBIR application.
Differentiators
- Zero vendor conflict. No AI vendor, platform, or reseller relationships to disclose or defend.
- Continuous, not point-in-time. Assurance stays current between reviews without a new procurement.
- Cryptographically defensible. Findings are hash-linked and signed — provable as unaltered.
- Classified-capable. Operates inside air-gapped facilities.
- Named human accountability. No certificate exists without a qualified auditor's signature.
- Multi-jurisdiction. 47+ regimes encoded; one assessment speaks to all of them.
Frameworks Assessed
- NIST AI RMF 1.0; NIST AI 600-1 GenAI Profile
- EO 14110 and successor federal AI policy; OMB AI memoranda
- NIST SP 800-37 RMF; SP 800-53; FedRAMP boundaries
- CMMC 2.0 and NIST SP 800-171
- DoD Responsible AI Ethical Principles
- Section 508; federal civil-rights and algorithmic accountability
- ISO/IEC 42001, 23894, 17020, 27001
- EU AI Act; Colorado ADMT Act; NYC Local Law 144
- SR 11-7; ECOA; HIPAA / HITECH
- Singapore IMDA agentic AI framework; OWASP Top 10 for Agentic AI
Certifications & Standing
- SAM.gov ACTIVE, all award types, since 3 Jun 2026
- CAGE Code 20JV1; UEI CB76ENDLMUC9
- Delaware corporation, established 2026
- ISO/IEC 17020 Type A — in process, not yet granted
- ISO/IEC 42001 — in process, not yet granted
- 3 U.S. patent applications pending — none granted
- No socio-economic set-aside designation claimed
- Name
- Ovi Pinzaru
- Title
- Founding Partner & CEO
- Phone
- (943) 243-0151
- Address
- 300 Colonial Center Pkwy
Roswell, GA 30076