🏛️ I architect SoC security at AMD — hardware-rooted security across x86, ARM, and RISC-V silicon
🔐 I own trust boundaries, security requirements, and threat models — the decisions that land long before anything reaches RTL
🔒 I build hardware-rooted secure sandboxes for LLMs — silicon trust models applied to what an autonomous agent is allowed to touch
🧪 Part-time hobbyist and open-source AI tool creator — everything on this profile is built outside the day job
🛡️ Founder of AlphaZede: containment, governance, and evidence infrastructure for AI coding agents
💬 Ask me about hardware roots of trust, isolation architecture, confidential computing, or agent containment
⚡ Fun fact: 6.5k+ contributions in the last year — the overwhelming majority behind private repos

_{alphazede.com is currently being redesigned.}

Bearing Lite — a skills-first Agent Plugin for planning, routing, bounded execution, and independent review. It preserves human authority and reviewable handoffs without shipping a CLI, MCP server, browser control room, or hidden runtime state.

npm pack @alphazede/bearing-lite

Install it through a compatible Agent Plugin client. The portable package is @alphazede/bearing-lite.

BRAN — local repository intelligence. Deterministic scanning, focused evidence packets, validation, and offline browsing. Headless for scripts and agent workflows, with an optional terminal interface for reading. Scanning and validation need no agent account.

Everything runs on your machine. No telemetry, no hosted middleman.

Defining trust boundaries, security requirements, and threat models for hardware-rooted security — and reviewing what silicon does with them. The work spans hardware root of trust, isolation and zero trust architecture, confidential computing, and specifications realized in SystemVerilog and Verilog.

_{Most of my work ships from private repositories — the totals above include those contributions.}

_{Building boundaries for autonomous systems · Austin, TX}