AI Architect · Austin, Texas · Website · Email · LinkedIn · X
I design agentic systems where a model can propose and cannot grant itself permission. Financial services trained the consequence. Founding trained the ownership. The public repos are the parts a reviewer can inspect.
Current: Founder, Apex AI|ML. Prior: Minotaur Consulting · Office of the CIO · Wall Street.
| Order | Repository | What it proves | Where it stops |
|---|---|---|---|
| 01 | Agent Foundry | Capability, identity, and runtime grants are separate. 399 tests. AWS DEV. | Not a customer-production system. |
| 02 | Monster Heavy | Durable paper execution survives retry, concurrency, and worker death. | No broker. No real-money path. |
| 03 | AI-Ready Data Platform | Three specialists. Claims checked against bounded facts. | Fixture timing is not a live-model speedup. |
| 04 | Research desk | Equities and energy agents draft. A clerk accepts. | A memo is not an order. |
| 05 | Investment gems | Local six-name synthetic screen across equities and energy exposure. | A watchlist is not a ticket. |
| 06 | Paper trading floor | Three stdio MCP servers expose fixture, risk, and paper OMS tools. | Venue tool is denied. |
| 07 | BALLAST | Explainable supply-chain risk, what-if shocks, and a disruptive-action approval gate. | Synthetic energy and semiconductor scenarios; simulated execution. |
| 08 | AEGIS Evidence | Replayable SOC assurance workpaper, human acceptance, open gaps, and SHA-256-verifiable archives. | Indicative framework mappings; not certification or a legal compliance determination. |
These eight projects show depth in governed AI and breadth across capital markets, energy, supply chain, and cybersecurity assurance. BALLAST makes operational trade-offs visible; AEGIS Evidence makes governance claims and unresolved control gaps inspectable.
Monster Light and Monster Desk remain supporting paper-boundary examples alongside Monster Heavy.
A recommendation is not permission. Approval is bound to the exact action. Current evidence is checked again at execution. If the control cannot be shown, the claim is not made.
Three static HTML fixture views, not live feeds or observed market data. Three desks, one limit. Research can write a memo. The screen can write a watchlist. Only the paper floor can write a fill, and that fill has no venue. Its three MCP servers are verified over stdio; the research and screening examples are deterministic.
- Research desk dashboard rejects a stale curve and an unsourced headline.
- Investment gems dashboard shows the hurdle and the invalidation.
- Paper floor dashboard shows the ERCOT paper fill and the denied venue tool.
Evidence scope and confidentiality
Confidential client material is not published. Public FinOps repositories contain modeled scenarios with documented formulas and assumptions; their dollar figures are not delivered client savings. Employment outcomes are reported separately on my resume and should not be inferred from these demonstrations.
Senior seats in AI architecture, enterprise AI, cloud and AI platforms, technical consulting, and technical chief of staff. Austin, Dallas, Houston, or San Antonio. In office or hybrid.
Close enough to build. Senior enough to say no.