Capability deep-dive
Software & ML development
How we deliver, the governed four-layer AI architecture, our stack, and compliance posture for federal software and applied AI programs.
Explore the capability →What this is
Viceroy NM's Software & Applied AI division builds secure, government-grade enterprise software and applied AI systems for federal agencies, defense organizations, and national laboratories. It is delivered with Trunnion AI, the group's enterprise software company (both Viceroy NM and Trunnion AI are wholly owned Duskbridge subsidiaries), and covers custom software, workflow automation, and governed multi-agent AI built to FedRAMP-aligned architecture and NIST 800-53 / CMMC Level 2 design alignment, deployable from commercial cloud to fully air-gapped environments.
What we build
Custom, secure software and workflow automation delivered on government timelines with clean documentation and full audit readiness. Modernization without rip-and-replace risk, built on a production-grade stack and deployed into your existing environment.
Governed multi-agent systems that execute real workflows, not demos. Every action flows through a deterministic policy engine, every decision touches an audit log, and every autonomous step has a human checkpoint. Reliability and oversight before raw capability.
Explore
Capability deep-dive
How we deliver, the governed four-layer AI architecture, our stack, and compliance posture for federal software and applied AI programs.
Explore the capability →The platform
The LLM-agnostic multi-agent orchestration platform behind the portfolio: policy engine, audit trails, and human-in-the-loop controls.
Visit trunnion.ai →The portfolio
Ten governed Trunnion AI products tuned per vertical, from procurement automation to defense agent frameworks, on one foundation.
View the portfolio →Prefer to read first? See ourGovernment AI insights on evaluation, oversight, and right-sized models.
Common questions
Viceroy NM builds custom enterprise software, workflow automation, and multi-agent AI systems for government and defense. The architecture is designed for isolated or classified-environment deployment only where a program supplies the accredited environment and appropriately cleared personnel; Viceroy NM does not represent that it holds a facility clearance or cleared-personnel status.
Federal agencies, defense primes, and DOE and NNSA national laboratories including Sandia National Laboratories and Los Alamos National Laboratory. Systems support supply-chain vendor qualification, manufacturing oversight, procurement automation, and secure, auditable AI in high-consequence missions.
Commercial chatbots answer prompts. Viceroy NM builds governed agents that execute workflows through a deterministic policy engine, with every autonomous step gated by a human checkpoint and recorded in a tamper-evident audit log. Reliability, oversight, and auditability come before raw model capability.
The architecture is designed for NIST 800-53 and CMMC Level 2 control mapping and for isolated deployment. Use in a classified program requires a program-provided accredited environment, appropriately cleared personnel, an approved security boundary, and written authorization. Viceroy NM does not represent that it holds an FCL, personnel clearances, CMMC certification, or FedRAMP authorization.
Evaluating the AI offering? Read the plain-languageAI Capabilities & Limitationspage for intended use, deployment contexts, the human-review expectation, and known limitations, and the Security & Trustpage for what is held, in progress, and not held today.