Full-Stack & AI Engineer
Responsible, Governed AI · Human-in-the-Loop · Source-Verified Knowledge Systems
I'm a recent graduate focused on responsible, governed AI — human-in-the-loop systems with verified source-grounding and anti-fabrication discipline, backed by the secure full-stack engineering to ship them. I turn unstructured information into structured, source-verified, queryable knowledge. Everything below is something you can open, run, or read.
I work both sides of the decision: the analytics that turn fragmented data into answers, and the secure engineering that puts those answers into production.
My path into this wasn't a straight line, and that's where the value is. It began in defensive security: an NSA/NSF-funded cybersecurity camp at the University of San Diego, then a cyber-security internship with the City of San Diego. That foundation — thinking about how a system fails before thinking about how it ships — never left.
At San Diego State I layered a Management Information Systems degree with a Computer Science minor and an interdisciplinary Honors track. MIS taught me how systems serve a business, Computer Science gave me the depth to build them, and the Honors work kept both grounded in how real people actually use what I make. A 250-hour design internship in Rome added the part most engineers skip: how a system communicates its purpose is part of the engineering, not a coat of paint.
From there the work moved into data and AI. As an independent consultant I built governed-AI knowledge systems for private clients — turning unstructured professional expertise into searchable, source-verified, traceable research and operations libraries, with a human reviewer in control of every decision. A CDC public-health analytics project sharpened the statistical side — model comparison, honest evaluation, knowing when a headline number is misleading. A commercial-operations and AI-solutions internship at Celltrion put it to work inside a regulated environment: Power BI decision-support dashboards and an Azure OpenAI integration plan with human-in-the-loop review. Graduate-level reinforcement-learning coursework taught me to trust a design only after it survives adversarial validation.
What ties it together is one pattern I keep applying: take fragmented, unstructured information and turn it into governed, source-verified, queryable knowledge that stays dependable under real, messy use — with a human in the loop and an anti-fabrication discipline that refuses to invent what it can't cite. I'm equally comfortable surfacing the insight and architecting the governance layer that makes it trustworthy — and I hold all of it to a single rule: everything you see here, you can open, run, or read.
Every status is exact — never rounded up.
Security here means specific, checkable things — not a credential I don't hold. This very site ships strict HTTP security headers and a per-app Content-Security-Policy, keeps zero secrets in the client bundle, and is deliberately inert (no database, no auth, no custom server route) to remove whole classes of vulnerability rather than patch them. Where input does exist, it's validated at the boundary.
I wire LLMs into real workflows the way you'd run a system you're accountable for: outputs grounded in verified sources rather than left to fabricate, model output treated as untrusted, an explicit tool allowlist instead of arbitrary calls, a human in the loop for anything irreversible, and a design that keeps behavior predictable under real, messy use — not just in the demo.
Front-end through deploy: typed React / Next.js interfaces, automated pipelines (GitHub Actions running lint, tests, build, and secret scanning), and supply-chain hygiene — pinned dependencies, lockfile-enforced installs. The projects below show the range end to end, not just the front of it.
Validated Research LibraryA governed-AI knowledge system turning unstructured expertise into a source-verified, queryable library.
Governed Operations SystemA queryable operations knowledge library for a small-business client, with a human committing every change.
AlignFlowA multi-source data-quality / master-data-management engine for reconciling records into one trusted view.
CDC Diabetes-Risk Analytics POCAn academic ML proof-of-concept classifying three-class diabetes risk on a large CDC public-health dataset.
VisionAir
mediCalmA beta, safety-bounded health PWA.
VocaLatticeA local-first audio-ML pipeline for vocal processing.
Independent AI Consultant (2025 – Present) — building governed, source-grounded AI knowledge systems for private clients in two domains.
Commercial Operations, Data Analytics & AI Solutions Intern (summer 2025) — built the analytics and decision-support layer for a commercial-operations team.
UX/UI Design Intern — a 250-hour international internship in Rome, Italy (summer 2023).
Cyber Security Intern (2022) — defensive-security fundamentals and cohort learning operations.
Education
B.S. Management Information Systems, CS minor — honors, cum laude. San Diego State University, Weber Honors College. Graduated May 2026.
By conversation
I've also built and run real, governed LLM systems for two paid clients in very different domains — an independent physical-therapy professional and a culinary operation. The systems I engineer turn each client's unstructured expertise into a validated, queryable knowledge library, with source-grounded outputs and a human reviewer in control of every decision. The system architecture is my work; the clients' content is theirs — so these are described here, never quantified or linked. I'm glad to walk through the architecture and the reliability engineering in a conversation.
The simplest path is best. A first message can just be “let's talk” — tell me the role and what you're building, and that's plenty to start.