A recap of my DAMA Philadelphia session on LLM05:2026 data and model poisoning, trust, RAG integrity, and the controls that keep AI-accessible data from quietly becoming an attack surface.

Agentic AI security needs a safe, repeatable target where attacks against tools, memory, retrieval, identity, and other agents produce measurable evidence. That is the idea behind AgenticGoat.

The Intelligent Control Process embeds governance, AI usage standards, human review, verification, and traceability into the full delivery lifecycle.

A practical look at designing a production-oriented agent platform with real user interfaces, multi-agent orchestration, durable run state, MCP-based tool integration, and cost-aware model routing.

Designing a React-based IAC Workbench that lets architects define extension templates and builders generate schema-valid HACTU8 test, tool, and attack packages.

Developing a pluggable extension system for the IAC (Intelligence Assurance Center) Prototype.

A recap of my experience at Bsides Philly 2025 — highlights, insights, and what comes next.

How CAG and Azure’s Model Router create a recommended architecture for modern ITSM, ticket orchestration, and automated support workflows.

A deep dive into the architecture pattern that dominated AI agent deployments in 2025: one immutable agent binary, externalized context, and proximity-based execution.

After two decades debugging production systems across defense, healthcare, financial services, and gaming, I've learned that "it works most of the time" isn't good enough. Here's how to make LLM-driven skills behave.

Each decade replaces friction with abstraction. AI is no different. This is not a bubble; it’s the next architectural leap.

Over the past few months, I’ve led the design and development of a Proof of Concept (PoC) engagement co-sponsored by Microsoft and a trusted enterprise client.