Available · SF or Seattle · US Remote

Humanistic principles first.

Production AI
Work that matters.

Founder of PREVERA Guardian+AI — USPTO patent pending. Designer of À la Carte — a Snorkel Flow / Foundry showcase. DOI-published RAG pipeline on Hugging Face. I build production AI on Claude's tool-use API, V-JEPA 2.1 world models, and twenty-one years of working at the edge of clinical reality.

View work → Get in touch
1
USPTO patent
pending
1
DOI-published
RAG pipeline
5
AI systems
shipped
1
Public design
case study
01 — About

From bedside to build.

I build human-in-the-loop AI for users the default system was failing — first as a frontline clinician, then as a special education teacher, now as a technical founder and designer. The methodology is the same across every chapter: design the system that lets a real person do something hard, with humans in the loop and the stakes visible. AI is the first toolkit that's caught up with the ambition.

Today I ship end-to-end: agent architectures on Claude's tool-use API, V-JEPA 2.1 world models on NVIDIA Jetson Orin, full-stack RAG pipelines, design-engineering case studies. One USPTO patent-pending clinical AI product. One DOI-published RAG pipeline. One public Snorkel Flow / Foundry showcase. Range and depth, by deliberate design.

I.

Clinical AI

Twenty-one years of clinical, education, and admissions work. Anesthesia, psychiatry, special education — domain depth that informs every system I design.

II.

Agentic AI

Production agents on Claude's tool-use API and MCP. ReAct orchestration, multi-step planning, evaluation harnesses with adversarial test sets.

III.

Design-engineering

Working prototypes that connect to real systems. Iterative bake-offs with Claude and Cursor. Visual and interaction craft pushed until it ships.

IV.

Responsible AI

Fairness audits, bias detection, SHAP interpretability. Programmatic labeling and weak supervision as production patterns. Human-in-the-loop refinement.

02 — Selected Work

Flagship systems first.

Patent-pending clinical AI, public design-engineering case studies, DOI-published research, and reproducible open-source builds — each solving a specific problem at the intersection of expert domain knowledge and machine intelligence.

01 — Public design case study May 2026 Live

Snorkel Flow / Foundry showcase · solo design and build

À la Carte

A reading-group companion app — and a working argument that the typed-component pattern from frontier-lab game-playing research generalizes outside of games. Six recipe cards, five lifecycle states, one canonical card catalog that fills session by session as filed cards accumulate. Built end-to-end in a deliberate bake-off using Claude and Cursor. Demonstrates the AutoHarness / Code-World-Models pattern applied to a non-clinical product surface.

HTML/CSS/JS Claude Cursor React Snorkel Flow Foundry
View case study
02 — Patent-pending Mar 2026 USPTO PGA-2026-001

V-JEPA 2.1 + 3D LIDAR · sole-inventor patent · Edge GPU

PREVERA Guardian+AI

Dual-path agentic fall-prediction pipeline deployed to NVIDIA Jetson Orin. Fast path: under 200ms fall detection. Prediction path: forecasts pre-fall motion 2–3 seconds before ground impact via V-JEPA 2.1 encoder/predictor over BEV depth maps. Sensor fusion with LIDAR geometric features, pressure mats, and clinical context. HL7 FHIR R4 integration. Fairness audited across mobility and body-type subgroups.

V-JEPA 2.1 Jetson Orin Python PyTorch HL7 FHIR SHAP
preveraguard.com
03 — Public agent May 2026 HF Spaces · Agents

Research & Snorkel-Reading-Group copilot

Remy V1

Agent for research and reading-group support, running live on Hugging Face. Built for my own use case as an active participant in research-community reading groups — Snorkel's included. Public, runnable, continuously updated. The kind of community-facing tool an AI Advocate ships.

Claude API Agents Python Gradio Hugging Face
Try Remy on Hugging Face
04 — Published research 2025 DOI 10.57967/hf/8101

Full-stack RAG · clinical decision support

FetchMerck AI

End-to-end RAG pipeline published on Hugging Face with DOI: ingestion, chunking, embeddings, semantic retrieval, re-ranking, prompt engineering, deployment. Built LLM-as-a-Judge scoring rubrics across multiple clinical query classes — the evaluation infrastructure enterprise AI engagements need to prove value realization.

Mistral-7B LangChain RAG Hugging Face Python
Hugging Face profile
05 — Full ML lifecycle 2025 Apache 2.0 · HF

scikit-learn pipeline · retail sales forecasting

SuperKart Sales Predictor

scikit-learn Pipeline (tuned Random Forest) for per-product per-store retail sales forecasting across Tier 1–3 cities. Held-out test set: RMSE 277.28, R² 0.9326, MAPE 3.91%. Deployed end-to-end on Hugging Face — model + Streamlit frontend Space + Flask backend Space, all public, all reproducible from the model card.

scikit-learn Random Forest Streamlit Flask Hugging Face Spaces
Model on Hugging Face

Also built — earlier work & explorations

03 — Credentials

The training behind the work.

Graduate research methodology, frontier AI training, and cloud fundamentals — the foundation for every system I ship.

UT Austin · McCombs

Post Graduate Program in AI/ML

Great Learning · McCombs School of Business · GPA 3.87

Completed May 2026
Saybrook University

PhD ABD — Integrative Mind-Body Medicine

Classic Grounded Theory methodology

Dissertation stage
Chaminade University of Honolulu

MS Counseling Psychology · BA Psychology

BA Magna Cum Laude

Completed
Anthropic

Anthropic Academy — 13/13 Courses

Claude API, prompt engineering, agentic systems

Certified
AMD

AI Developer Path — Agents 101 & 201

Agent architectures on hardware-accelerated inference

Certified
Microsoft

Azure AZ-900 & AI-900

Cloud fundamentals · ML, computer vision, NLP, generative AI on Azure

Certified
Databricks

Advanced MLOps & SQL Analytics

Machine learning operations and SQL analytics badges

Certified
AAPB · BCIA

Biofeedback Certification

Biofeedback Certification International Alliance · applied psychophysiology

Certified
USPTO

Patent Filing — Sole Inventor

PGA-2026-001 · LIDAR + V-JEPA 2.1 fall-prediction architecture

Provisional filed
04 — Journey

A non-linear path with through-line.

Every role taught something that compounds in the AI work: clinical reasoning under pressure, mixed-methods research, multi-stakeholder program delivery, and consultative communication with expert and non-expert audiences alike.

2024 — Present

Technical Founder · PREVERA Guardian+AI

Independent · Remote

Shipped five AI applications, zero-to-one, with one USPTO provisional patent filed as sole inventor. Built on Claude's tool-use API. Full ownership: research, architecture, deployment, evaluation, patent prosecution. Concurrent advisory engagement with The Gerontechnology Group on ethical AI for aging populations. Designed and shipped À la Carte (May 2026) as a public design-engineering showcase.

2019 — 2024

Special Education Teacher

Hawai'i Department of Education

Designed data-driven experimental interventions for 20+ students per year — applied mixed-methods experimental research with iterative calibration under tight timelines and real stakes. Coordinated cross-functional IEP teams as the portfolio-based delivery analog of professional services. The literal lifecycle-state-refinement loop that À la Carte's architecture is modeled on.

2016 — 2019

Associate Director of Admission & Adjunct Faculty

Hawai'i Pacific University

Consultative delivery at enterprise scale: understood prospective-student objectives, designed academic solutions across programs, navigated multi-stakeholder decision processes, and drove outcomes against quarterly enrollment targets. Adjunct psychology faculty — taught research methods and delivered technical presentations to diverse audiences.

2011 — 2016

Special Education Teacher (first stint)

Hawai'i Department of Education

First five-year stretch of IEP portfolio delivery across diverse classroom contexts. Learned to coordinate cross-functional teams under legally binding delivery targets.

2003 — 2011

Anesthesia & Psychiatric Technician

Queen's Medical Center · Level I Trauma Center

Eight years of frontline observation of clinical reasoning under uncertainty during information-dense decisions. Ground-truth domain expertise in perioperative workflows, malignant hyperthermia response, and psychiatric stabilization — the clinical foundation anchoring every AI system I now build.

05 — Stack

The toolkit.

A practitioner's stack — verified through the patents, published research, and shipped case study. Not resume decoration.

Languages

  • Python
  • TypeScript / JavaScript
  • SQL
  • R · HTML/CSS

ML / AI

  • PyTorch
  • Hugging Face Transformers
  • scikit-learn
  • XGBoost · SHAP
  • V-JEPA 2.1

LLMs & agents

  • Claude API (tool-use)
  • MCP
  • ReAct agents
  • LangChain
  • RAG · Mistral-7B

Build & design

  • Cursor
  • React
  • FastAPI · Streamlit
  • NVIDIA Jetson Orin
  • Databricks · Azure · HL7 FHIR R4
06 — Contact

Let's build something that matters.

Open to Applied AI Engineer, Research Engineer, Staff / Senior Product Designer, Solutions Architect, and AI Engagement roles at frontier AI labs and healthcare AI companies. Remote, or relocating to SF or Seattle.

Also open to advisory and consulting engagements — especially at the intersection of clinical AI, evaluation frameworks, and human-in-the-loop product design.