My Work: From Principled Research to Practical Implementation
Here is a selection of my major projects. Each is an exploration in building systems that are not just intelligent, but also transparent, auditable, and accountable by design. Click on any project to see a detailed case study of the research, architecture, and outcomes.
Controlled AI Execution
Architect
An AI system for technical task execution under strict human control.
Problem: Most autonomous agents act as black boxes, making it difficult to inspect their logic, stop their actions, or reverse their mistakes.
Approach: Built an execution loop that gates and records actions, requires testing before capability promotion, and keeps higher-risk operations under explicit human control.
Outcome: A dependable system that can perform technical tasks while remaining observable, reviewable, and accountable to the operator.
Behavioral Analytics
NeuroTrace
A framework for analyzing how AI behavior changes over time.
Problem: AI personalization is often judged through intuition, making it hard to measure how an assistant actually adapts across many sessions.
Approach: Created a deterministic pipeline that turns conversation logs into structured behavioral data, mapping patterns such as planning, memory, feedback, and reflection.
Outcome: A repeatable way to inspect behavioral change over time and discuss personalization with evidence instead of guesswork.
Content Infrastructure
MD Grid Engine
A content engine for building fast, file-first websites. It also serves as the UI foundation for Architect, where the same system is used to present plans, execution traces, and decision outcomes in a structured way.
Problem: Traditional CMS platforms are often too rigid or heavy for technical sites that need structured content and more complex system views.
Approach: Engineered a custom Next.js engine that treats content as code, transforming Markdown and MDX into responsive layouts through a centralized component registry.
Outcome: A fast, maintainable system that powers this portfolio today and provides a UI foundation for future Architect interfaces.
Controlled Writing Workflow
Writing Agent
A multi-step writing system for controlled rewriting and tone management.
Problem: Most AI writing tools behave like one-shot black boxes, making it hard to judge faithfulness, tone, or overall quality in higher-stakes communication.
Approach: Built a structured rewrite pipeline that plans, drafts, evaluates, corrects, and polishes output instead of relying on a single prompt call.
Outcome: A more inspectable writing system with built-in quality checks, tone controls, and traceable output for real communication work.
Governed Retrieval Architecture
CORA
An architecture concept for governed retrieval and traceable knowledge work.
Problem: Most retrieval systems return answers with limited visibility into routing, evidence quality, verification, or privacy handling.
Approach: Designed CORA as an evidence-first architecture with structured routing, verification layers, provenance tracking, and observability built into the workflow.
Outcome: A blueprint for retrieval systems that are more inspectable, evidence-backed, and easier to trust in serious knowledge work.
Stateful Support Workflow
Communication Agent
A stateful email triage and response system for support workflows.
Problem: Email automation often stops at summarizing or drafting, without managing ticket state, approval steps, or secure operational handling.
Approach: Built a backend-first system that connects to Gmail, manages ticket states, supports model-assisted triage and drafting, and keeps the workflow under operator control.
Outcome: A structured support system with secure ingestion, persistent state, and human-reviewed responses across the full email lifecycle.
Roadmap
Current Focus & Future Work
I’m currently building a connected portfolio of governed AI systems, observability-first tools, and evidence-backed agent workflows. The direction is moving from isolated prototypes toward a more unified stack where retrieval, governance, writing, analytics, and interface layers can work together as one coherent system.
Now Available:
- MD Grid Engine - the content and interface engine behind this portfolio
- NeuroTrace - the core cognitive analytics framework for tracing assistant behavior and interaction patterns
- Writing Agent - an auditable, self-correcting agent workflow for controlled text rewriting and tone transformation
In Progress:
- CORA - a governed, evidence-first retrieval and orchestration system focused on provenance, verification, and trustworthy knowledge workflows
Up Next:
- Connecting Architect, CORA, and the MD Grid Engine into live visual demos of governed AI workflows
- Extending observability and evaluation across projects so decisions, evidence, and failure modes stay visible by design
- Turning more blueprint-stage systems into runnable demos, release-ready repositories, and public technical documentation