I turn ambiguous requirements into systems that run. I have been working on connecting data pipelines, backend APIs, applied AI, and product experience. I've built healthcare monitoring infrastructure, automated semiconductor data workflows, and taken my own product from concept to a live beta.

USC M.S. Spatial Data Science, expected December 2026 · Los Angeles · Open to 2027 full-time roles

Portfolio · LinkedIn · Email · DuoSky live

Explore my work: Product & UX — DuoSky · Data systems — Health Pipeline · Agent workflows — AIvestor

A live, trilingual BaZi × Western astrology product that turns birth details into layered, explainable readings and shareable birthplace portraits.

• Owned product scope, architecture, UX, validation, and release decisions; shipped on Cloudflare Workers.
• Designed progressive input and reading depth: start with a birth date, add time and place for precision, and explicitly leave time-dependent conclusions open when birth time is unknown.
• Built interactive chart explanations and side-by-side evidence comparisons, plus shareable portraits that omit names, birth dates, and raw birth times.
• Built birthplace resolution covering 170,949 places across 246 countries and territories, with historical timezone handling and 256 lazy-loaded data shards.
• Added GIS-based city portraits and geometry fallbacks; built regression coverage across 720 chart profiles and 5,760 yearly cases. These are test-coverage counts.

JavaScript · Python · Cloudflare Workers · GIS · Automated QA
Try DuoSky

A public, synthetic-data prototype demonstrating device-data validation, deduplication, time-series aggregation, layered anomaly detection, and prioritized alerts. Includes sensor faults, duplicates, gaps, and calibration drift.

Python · pandas · NumPy · SQLite

Built User Preference and Portfolio Optimization agents in a five-person multi-agent investment-research project, connecting macroeconomic data, market analysis, and personalized portfolio workflows with structured JSON/Markdown reporting.

Python · LangChain · LLM APIs · FRED · Agent orchestration

I also build spatial ML workflows: H3-based Los Angeles crash-hotspot forecasting, time-based model evaluation, and ArcGIS prediction maps.

xHealth Group — Summer AI Engineering Intern · June–August 2026

Owned an always-on wearable-monitoring platform from problem framing to GCP deployment: self-service OAuth, ingestion of 6+ biometric streams, canonical data models, scheduled processing, and explainable care-team alerts.

• Migrated SQLite to PostgreSQL with zero behavior drift across 398,802 synthetic evaluation readings and 166 alerts, backed by row-level parity and a frozen evaluation fingerprint.
• Built versioned JSON contracts for a four-project ecosystem, encrypted token storage, least-privilege access, and 300+ automated tests, including checks inside the deployed container images.
• Built the AI-CRM analytics and decision loop: scheduling → measurement → enrollment attribution → next-best actions → feedback, with organization-scoped APIs and small-cell privacy controls.
• Researched personalized anomaly detection with cross-signal corroboration and conformal scoring, using patient-level splits and bootstrap confidence intervals. This was research evaluation, not a claim of clinical validation.

UNIM Innovation — Data Science Intern · June–August 2025

• Built reproducible Python/SQL workflows over 50K+ wafer-test samples and 270+ weekly batches, supporting manufacturing analysis and forecasting.
• Automated raw test-archive parsing into ERP-ready outputs, preserving lot, wafer, and BIN granularity; reduced manual processing by ~90% and made the workflow usable by Finance and Operations.

IDXExchange — Data Scientist (Internship) · January–April 2026

Led a five-model comparison for California home-price prediction, using geospatial features and forward holdout validation; selected XGBoost with R² = 0.8843.

I use Codex, Claude Code, and Cursor to accelerate implementation, with explicit acceptance criteria, independent review, regression checks, and human release decisions.

Let's build something useful. I'm looking for roles in Forward Deployed Engineering, Applied Data Science, and Associate Product Management (APM) / AI Product starting in 2027. I'm also open to Data Engineering and applied AI roles focused on integrations and delivery, with flexibility for U.S. relocation and travel.

Outside engineering, I make R&B, hip-hop, and soul as JEdoubleF911. Explore both sides →