Core competencies
- AI Readiness Assessment: use-case inventory and risk tiering, readiness scoring, phased roadmap. Typically 3–4 weeks.
- Assured Build: retrieval-augmented and agentic AI integrated with existing systems, with human review gates, an accuracy baseline, and an evidence package built alongside the system. First increment typically 8–12 weeks.
- AI Verification and Baseline: black-box and adversarial testing (prompt injection, data disclosure, ungrounded output, excessive agency), subgroup testing, and a crosswalk for the Risk Management Framework (RMF) authorization package. Typically 4–6 weeks per system.
- Continuous Monitoring and Drift Detection: re-evaluation on any model, prompt, or retrieval change; quarterly posture reporting.
- AI Governance: AI usage policy, risk-tiering criteria, and maturity roadmaps. Typically 2–3 weeks.
Differentiators
- Evidence from day one: testing, baselines, and audit trail produced as the system is built, not reconstructed for review.
- Inside the authorization boundary: work in the customer's or prime contractor's environment, under NIST SP 800-171 for CUI, DISA Cloud Computing SRG impact levels, and HIPAA for protected health information (PHI). No CUI is stored on company systems.
- Model-agnostic: policy-approved models on Azure or AWS, or local models on isolated infrastructure.
- Everything delivered: source code, services, test harnesses, assurance tooling, evidence packages, and monitoring configurations, under the contract’s data-rights clauses (DFARS 252.227-7013/-7014 or FAR 52.227-14); no proprietary lock-in.
- Framework-aligned: NIST AI RMF and Generative AI Profile (AI 600-1), ISO/IEC 42001, OMB M-25-21, RMF (DoDI 8510.01; NIST SP 800-37).
Principal’s experience
- Chief Engineer, Medical Community of Interest, DISA: delivered DoD/VA's first integrated healthcare data network, serving 15M+ service members and veterans.
- Chief Data Scientist, GSA FAS: key contributor to GSA's first AI governance policy; predictive models identifying $170M+ in sales opportunities; classification models across $3.5B in vendor product analysis.
- Chief Data Architect, GSA FAS: designed and delivered the FAS Enterprise Data Architecture, integrating 32 data sources into AI-ready AWS environments; led a team of 14.
Experience of the firm's principal in prior roles; not past performance of Analytica Logica LLC.