Selected Builds

Production systems for data, apps, analytics, and AI.

I build the machinery behind modern data products: cloud platforms, pipelines, APIs, dashboards, internal tools, governance workflows, and AI-assisted applications. My deepest current work is in AWS and Azure, with earlier GCP experience in analytics and warehouse modernization.

Multi-cloud background

AWS + Azure + GCP

Data products at scale

Millions of rows/day

AI with guardrails

Agents + rules

Production discipline

IaC + CI/CD

Builder portfolio

These are public-safe summaries of production work and serious personal products. Source systems are described by domain where naming a specific platform would reveal more than the work needs to.

Personal flagship app: health tech, computer vision, and AWS

Range of Motion Tracker

A live telehealth application for physical therapists to run range-of-motion assessments, manage patients, assign exercise programs, and share progress with patients.

Next.js 16React 19TypeScriptMediaPipeAWS ChimeDynamoDBS3ClerkBedrock

View live app

What I Built

Browser-based pose detection with MediaPipe, real-time angle overlays, movement guides, and roughly 80 configured movements across major body regions.
Remote therapy sessions with AWS Chime, therapist and patient portals, role-based auth, patient management, program building, and completion logging.
Snapshot capture, S3 storage, DynamoDB-backed session data, PDF reports, progress dashboards, and optional Bedrock summaries behind a feature flag.

Why It Matters

Live public app with a full product surface: landing page, demo, admin, therapist, patient, local, remote, upload, and reporting flows.
Designed with healthcare-grade concerns in mind: role routing, audit logs, KMS-encrypted storage, CSP, and security documentation.
Built as a real SaaS-style product, not a demo: authentication, data model, cloud infrastructure, video, reporting, and patient-facing UX.

Shipped consumer iOS app: bilingual learning games

Little Learners Lab

A published App Store education app with bilingual English and Spanish games for early literacy, numeracy, colors, shapes, memory, puzzles, and motor skills.

SwiftUISpriteKitiOSLocalizationPhotosCameraAudio toolingApp Store

View App Store listing

What I Built

A SwiftUI and SpriteKit game library with 19 activities across letters, writing, numbers, math, colors, shapes, memory, mazes, photo puzzles, and sorting games.
Bilingual onboarding, settings, UI copy, game prompts, vocabulary, and audio flows for English and Spanish play.
Local-first family-photo and camera experiences, custom tracing paths, game-state sessions, celebration feedback, and audio-render tooling.

Why It Matters

Published through the App Store with a public listing, version history, education category placement, and cross-device iPhone/iPad support.
Designed for young children with no accounts, no ads, no in-app purchases, no analytics, no tracking, and no third-party SDKs.
Shows end-to-end product delivery outside enterprise work: concept, design, mobile engineering, localization, privacy posture, release, and iteration.

Telemetry, APIs, and cloud data platforms

Customer Telemetry Platform

A production platform for turning facility equipment readings into clean, customer-facing data products and internal analytics.

AWS LambdaS3EventBridgeGlueDelta LakeAurora PostgresAzure Event GridBicep

What I Built

Event-driven ingestion across landing zones, queues, serverless processing, and a Delta Lake warehouse.
A customer API backed by serverless Postgres, signed requests, batching, and aggregate query patterns.
A next-generation message delivery path with certificate-based authentication and customer-scoped topics.

Why It Matters

Processes millions of metric rows per day across a multi-site environment.
Designed for idempotency, late-arriving data, deduplication, compaction, and operational recovery.
Supported with runbooks and validation tooling for onboarding, troubleshooting, and customer questions.

Enterprise source systems into governed analytics

Enterprise Operations Lakehouse

A set of production integrations that turn messy operational platforms into reliable lakehouse tables, semantic models, dashboards, and embedded reports.

AWS GlueAppFlowMicrosoft FabricSparkDuckDBPower BIReactFastAPI

What I Built

Pipelines for enterprise IT, asset management, commissioning, construction, location, and financial data.
Schema-aware transformations for reference fields, choice values, currency, durations, timestamps, and location hierarchies.
Incremental sync patterns that balance fast change capture with full-refresh safeguards when source data gets stale.

Why It Matters

Reduced manual reporting friction by moving source-system data into governed analytical layers.
Built recoverable sync queues and parity between local development and production lakehouse execution.
Delivered embedded BI and custom dashboard views on top of the same trusted data.

Full-stack decision support with deterministic AI guardrails

Site Evaluation Platform

A geospatial decision-support application for screening infrastructure opportunities with maps, scoring, analyst workflows, and AI-assisted research.

ReactTypeScriptFastAPIDuckDBLeafletMicrosoft FabricAzure FoundryContainer Apps

What I Built

A spatial backend for sub-second radius queries across transmission, substation, generation, market, and regulatory data.
A deterministic scoring and reconciliation engine that compares submitted claims against ground-truth infrastructure data.
An AI layer that adds qualitative analyst considerations while keeping final scoring rules-based and auditable.

Why It Matters

Combines external research, geospatial analysis, source-system intake, and PDF reporting in one workflow.
Secured with enterprise identity, managed secrets, managed identity access, and health-verified CI/CD.
Designed as a production app, not a prototype: typed front end, modular API, tests, deployment gates, and audit logs.

Governance, access, diagnostics, and BI operations

Governed Azure Data Platform

Platform work that makes the modern data estate easier to operate: cataloging, stewardship, BI workspace automation, and permission diagnostics.

Microsoft FabricPower BI REST APIsPurviewMicrosoft GraphPythonManaged IdentityRBAC

What I Built

An enterprise data catalog with managed-identity scanning, PII classification, stewardship domains, and glossary-as-code patterns.
A Fabric and Power BI operations CLI for workspace inventory, refresh diagnostics, job history, and permission-chain tracing.
Automation for external-subscription audits, access review workflows, and long-running API scans with rate-limit handling.

Why It Matters

Turns platform knowledge into repeatable runbooks, scripts, and operating procedures.
Helps teams answer where data lives, who owns it, whether it is sensitive, and why a BI asset failed.
Connects governance to practical delivery instead of treating it as a separate compliance exercise.

Internal tools for engineering execution

Engineering Program Tracker

A lightweight internal app for managing data-source onboarding, delivery phases, documentation, and leadership status reporting.

ReactViteExpressSQLiteGitTypeScript

What I Built

A project tracker that models each build through develop, deploy, and document phases.
Drag-and-drop planning, task calendars, source-system status, and reporting views for recurring updates.
Repository scanning that shows real engineering activity next to the plan instead of relying only on manual status notes.

Why It Matters

Created a shared operating layer for a growing portfolio of data and application projects.
Reduced context switching by keeping project state, repo activity, and documentation close together.
Generated recurring status narratives that connect technical progress to leadership visibility.

I build the whole flow

Source systems, cloud infrastructure, lakehouse tables, APIs, front ends, reports, and runbooks all have to work together.

I keep the foundation honest

AI and analytics are only useful when the data, permissions, deployment process, and operating model can be trusted.

I ship tools people can use

The point is not architecture for its own sake. It is working software that helps operators, analysts, customers, and leaders make decisions.

The short version

I am a builder with deep data architecture roots. I turn operational complexity into production-grade tools that are reliable, understandable, and useful.

About me Read the field notes