Expert Focus: Fractional CISO · IT Director · AI Architect Core Value Proposition: Helping SMEs, PE-backed, and high-growth businesses build secure, scalable technology without sacrificing delivery pace. Expertise at the intersection of cyber security, IT leadership, cloud modernization, and practical AI adoption.

• Secure AI & Agentic Workflows: Designing operating models with mandatory human approval gates, audit trails, cost control, and defined security boundaries.
• AI Control Planes: Building governed orchestration layers — deterministic substrate (systemd, scripts, CI) first; LLM orchestration (Hermes) above; frontier models (Perplexity, Claude Code, Codex) for research, build, and review.
• Multi-Model Orchestration: Routing work across local (LM Studio), cloud (OpenRouter), and specialist (Perplexity for research, Codex for review) providers based on task class and hardware constraints.

• Cyber Security & Compliance: Cyber Essentials, ISO 27001, SOC 2, GDPR — practical readiness with demonstrable evidence, not shelfware.
• Risk Management: Transforming security from reactive tasks into owned programs with clear accountability and audit trails.
• M&A Due Diligence: Assessing technology estates, security posture, operational risk, and integration readiness.

• Modernization: Microsoft 365, Azure, hosting, networking — secure, observable, maintainable.
• Homelab & VPS: Production-grade infrastructure management with NetBox-as-truth, Prometheus/Loki monitoring, and AI-assisted ops.

Recent concentrated work across the portfolio demonstrates the operating model in production:

• 15 parallel Claude Code sessions deployed via git worktrees, each tackling a dedicated GitHub issue — all branches pushed and bulk-merged.
• 23 issues closed across a single session: runtime config changes (fallback provider reordering, auxiliary vision pinning), security hardening (CVE-2026-48710 verified patched, gateway auth audited loopback-only), documentation refresh, and issue consolidation.
• Observability stack deployed: Langfuse v2 (port 3001), Langfuse v3 worker + ClickHouse + MinIO + Redis, Promptfoo (port 3002), PGAdmin — 9 containers, all smoke-tested.
• Current-state documentation: v0.16.0, 32 skills, fallback chain reordered, model/provider health verified.
• Architecture defined: Deterministic-substrate-first policy (ADR-0020), governed agent activation blueprint, 5-profile model (operator/research/builder/reviewer/ops).

• 4 live systemd-based ops agents with deterministic evidence collection.
• NetBox as infrastructure source of truth, Prometheus/Loki monitoring.
• Repository canonicalized as the home and VPS infrastructure automation platform.

• hermes-community-edition — Fully free/nearly free Hermes deployment with local installers for Windows, Mac, Linux.
• hamnet — Public-facing HAMNET platform repo.
• hermes-enterprise-reference-architecture — Public reference companion for enterprise Hermes deployments.
• client-agent-appliance — Client-side agent deployment appliance.
• AI-Alpha-Radar — AI alpha news aggregator.
• FlowFinder — Agentic workflow finder from existing logs.

1. Security first, delivery always — Controls must reduce risk without stopping useful work.
2. Human approval at critical gates — Accountability remains explicit even with powerful automation.
3. Deterministic substrate first — Scripts, systemd, CI own execution; LLMs research, draft, review, route.
4. Evidence over aspiration — Governance leaves an auditable trail.
5. Pragmatism over theatre — Small, well-controlled changes over large, fragile programs.
6. Git as SSOT — All state changes committed before reporting.

• Organizations that have outgrown informal IT decision-making.
• Businesses under client/insurer security pressure.
• Companies adopting AI faster than their governance model safely supports.
• PE-backed and high-growth businesses needing fractional technology leadership.