# Memoriant, Inc.

**Verifiable AI for regulated industries.**

Building compliance-focused language models, training datasets, and evaluation tools for the defense industrial base. Our models are purpose-built on fully auditable base architectures with complete training data provenance, designed for air-gapped, on-premises deployment in environments handling Controlled Unclassified Information (CUI).

## What We Build

| Product | Description |
|---------|-------------|
| **Compliance Models** | Purpose-built language models calibrated for CMMC 2.0, NIST SP 800-171, NIST SP 800-53, HIPAA Security Rule, DFARS, and FedRAMP guidance |
| **Training Datasets** | Curated compliance Q&A from authoritative U.S. government publications, validated for regulatory currency |
| **CMMC Expert Platform** | AI-powered compliance platform for defense contractors: SSP generation, gap analysis, POA&M drafting, evidence export |

## Deployment Options

| Option | Description |
|--------|-------------|
| **Cloud Enclave** | FedRAMP-compatible cloud infrastructure. No hardware purchase required. |
| **On-Premises** | Dedicated customer hardware. Full data sovereignty, CUI remains on-site. |
| **Air-Gapped** | Physically disconnected. For CMMC Level 3 and classified-adjacent environments. |

All deployments use proprietary compliance AI with zero external API dependencies.

## Open Source Contributions

Active contributor to [NVIDIA garak](https://github.com/NVIDIA/garak) LLM vulnerability scanner. Building adversarial probes for testing LLMs deployed in regulated environments.

## Links

- **Platform:** [memoriant.ai](https://www.memoriant.ai)
- **Contact:** info@memoriant.ai
- **NVIDIA Inception Member**