Writing Hub
AI governance essays, reasoning systems notes, experiment logs, and technical writing across BioAI and engineering practice.
Project Topics
Building a Deterministic Governance Kernel: Separating Custody from Truth
CGF separates domain truth from custody mechanics, turning AI governance from Markdown/YAML policy language into deterministic, inspectable artifacts.
I Built an Ecosystem of 46 AI-Assisted Repos. Then I Realized It Might Be Eating Itself.
An ecosystem of 46 AI-assisted repos can become a closed loop. This article explores structural blind spots, self-validating toolchains, and the need for external validators to create intentional friction.
Your Agentic Stack Has Two Layers. It Needs Three.
Most agentic stacks cover tools and skills, but miss intent governance. Learn why a third layer is needed to stop AI drift, scope creep, and technically correct systems heading in the wrong direction.
Why Reasoning Models Die in Production (and the Test Harness I Ship Now)
Why reasoning models fail after deployment, what production drift looks like, and the test harness I use to catch fragile logic before itships
Implementing "Refusal-First" RAG: Why We Architected Our AI to Say 'I Don't Know'
Implementing refusal-first RAG means teaching AI to say “I don’t know.” This article explains evidence atomization, Slop Gates, and grounding checks that favor verifiable answers over plausible hallucinations.
LOGOS LawBinder: From Governed Reasoning to Audit-Grade Execution
This article explains how LOGOS v1.4.1 improves production AI reasoning with multi-engine orchestration, complexity-aware governance, and audit-friendly failure tracing.
LOGOS v1.4.1: Building Multi-Engine AI Reasoning You Can Actually Trust
LOGOS v1.4.1 is a multi-engine AI reasoning orchestrator that enforces consensus, traces failures, and applies governance profiles to reduce drift and make production reasoning more trustworthy.
LawBinder v1.3.0: Governance as a Kernel (Not a Guardrail)
LawBinder v1.3.0 shows how AI governance can run like a kernel, using deterministic Rust-based enforcement, replayable audit signatures, and bounded-latency policy checks in the critical path.
I’m Not Building AI Demos. I’m Building AI Audits (ASDP + Slop Gates)
Learn how ASDP and AI Slop Gates turn AI trust into auditable evidence, with CI/CD checks, drift policies, and governance artifacts that block weak, narrative-driven systems.
HRPO-X v1.0.1: from HRPO paper production-hardened runnable code
How HRPO-X turns the HRPO paper into production-hardened runnable code, with reproducible execution, governance checks, and deployment-ready structure.
Undo Beats IQ: Building Flamehaven as a Governed AI Runtime (Not a Prompt App)
Why governed AI runtimes outperform raw model IQ, and how Flamehaven uses undo, control surfaces, and fail-closed review to keep systems trustworthy.
Turning a Research Paper into a Runnable System
Turn a research paper into a runnable system. This article shows how HRPO’s core equations were implemented with bounded policy lag, KL rejection, and execution checks to test real-world fidelity.