Tampa Bay, Florida, USA
Senior full-stack engineer building enterprise API applications on C# / .NET, Angular, React, and SQL Server.
01 / About
I've spent 15+ years building enterprise API applications on C# / .NET and SQL Server, with Angular front ends over ASP.NET Core Web APIs. My SQL Server work runs deep: complex queries, stored procedures, indexing, execution plan analysis, and performance tuning under concurrent load.
Most recently I re-architected a legacy interval-polling Windows service into an event-driven RabbitMQ pipeline that handles 1M+ API requests around the clock. Before that I developed and maintained TOPS Software's proprietary accounting system, covering general ledger, receivables, payables, bank reconciliation, and financial reporting.
I build on SOLID principles and modern software engineering patterns, and I own my work end to end, from Sprint 0 and architecture review through delivery, code review, mentoring, and production root-cause analysis. Teams count on me to ship on schedule and to leave the codebase better than I found it. AI-assisted tools like Claude Code are part of my daily workflow.
See my code on GitHubSometimes it is the people no one can imagine anything of who do the things no one can imagine.
02 / For Employers
For companies hiring a senior full-stack engineer. These are the six areas I own on a product team, from the first design review through production support, and the standard I hold each one to.
Turning requirements from Product into scalable design blueprints. I research the problem space and the systems already in play, then break the work into epics, stories, and tasks with clear acceptance criteria.
From there I plan Sprint 0, lead the architecture review, and guide the development effort through delivery, documenting progress, decisions, and challenges so the whole team can see where things stand.
An engineered design is the keystone.
Schema design and T-SQL on SQL Server: stored procedures, indexing strategy, and execution plan analysis for queries that stay fast under load. On AWS, single-table DynamoDB models with keys and access patterns designed up front.
Data stays tight from the database to the browser: typed models in the API, versioned request and response contracts, and TypeScript interfaces that mirror them, so a change in one layer is caught at compile time rather than in production.
Maintainable, efficient code written to industry-standard conventions refined over 15 years, built on SOLID principles and modern software engineering patterns, and delivered on schedule.
Clean, consistent C# and TypeScript the next engineer can read without a tour, kept that way through analyzers, code review, and small, focused pull requests.
No over-engineering or anti-patterns.
Every feature ships with fully functional unit and integration tests in xUnit or NUnit, with Moq for mocking, covering the happy path, edge cases, and failure modes.
Tests run in the pipeline on every push, so refactoring stays safe and regressions surface before review rather than in production.
Tests are the guardrails.
CI/CD pipelines in Azure DevOps or GitHub Actions that build, test, and ship Docker images, from GitOps deployments with ArgoCD on Kubernetes down to lean deploy scripts for small serverless projects on AWS.
Environments are reproducible, rollbacks are a single step, and every release leaves a trail from commit to running service.
Production is instrumented with structured logging through Serilog, shipped to Seq, Datadog, or CloudWatch, with dashboards and alarms tuned so problems surface before customers notice.
When something does slip through, I switch on debug-level logging for the affected component, trace the request through the logs, run the root-cause analysis, and ship the fix with a test that keeps it from coming back.
03 / Portfolio
My entire career has gone into enterprise systems that belong to the companies I built them for, so that code stays in their private repositories. The projects below are my own. They are here to show how I design, write, test, and ship software when every decision is mine.
The tech behind this site, in four editions
The default edition of this site follows the KISS principle: keep it simple, and add nothing that is not earning its place. Plain HTML, CSS, and JavaScript with no framework, no build step, and no dependencies. The files live in a private S3 bucket that only CloudFront can read through Origin Access Control. CloudFront serves them worldwide over HTTPS with an ACM certificate, Route 53 points the domain at the distribution, and a GitHub Actions workflow authenticated with OIDC syncs the bucket and invalidates the cache on every push to main. No servers, no load balancer, no auto scaling group.
The Experience section is served by a C# API on Lambda and DynamoDB, with default HTML as a fallback.
You are reading the first edition, the reference the other three are built to match. The same page is rebuilt in Blazor WebAssembly, Angular, and React, each served from its own prefix of the same bucket, and every content change ships to all four in the same commit. The edition menu in the header switches between them.
View the source on GitHub
FHIR R4 patient explorer, Angular in front of C# on Lambda with the Firely SDK
A window onto a public FHIR R4 test server seeded with synthetic records. Nine operations under /api/fhir run as C# on .NET 10 on AWS Lambda, specified in the same OpenAPI document as the Experience API: five reads (patient search, summary, raw bundle, capability, and health) and four writes (create, replace, patch, and delete a demo patient) under a guard that keeps the seeded patient read-only and lets a visitor change only what their own session created. The explorer in front of them is an Angular app of its own at fhir.mvcprogrammer.com. The browser never talks to the sandbox; the function is the boundary, parsing every resource with the Firely SDK and mapping only the US Core fields to small DTOs.
Built to be polite to a shared sandbox. Every upstream call has a short timeout, retries with backoff, and a circuit breaker, and a failure comes back as an RFC 9457 problem the explorer turns into a plain message. Responses carry ETags and CloudFront caches them. No real patient data, nothing stored.
Photography studio, Angular on AWS
The website for LightSplitters Media, a fictional Tampa Bay photography studio I created as a portfolio piece: weddings, portraits, family albums, and archival prints. An Angular 22 workspace with standalone components and signals, prerendered to static HTML with no server. A second app in the same workspace is a template for client sites, so each couple's wedding site or family album lives at its own subdomain from one shared build plus a JSON file and a folder of photos. Masonry galleries with a native dialog lightbox, responsive WebP and JPEG images, a contact form that posts to a Lambda function sending through SES, and print orders through Stripe Payment Links. All of it served from a private S3 bucket through CloudFront, with a CloudFront Function routing each subdomain to its folder.
Touch screen kiosk, React + .NET on AWS
A touch-screen kiosk behind the window of a Clearwater Beach real estate office, running since 2010. Visitors pick sales or rentals over an aerial photo of Sand Key, browse live Stellar MLS listings and photos pulled through the Bridge Data Output API, and ask to be emailed details or called back from an on-screen keyboard. The 2026 rebuild is a React and TypeScript screen built with Vite, pixel-matched to the original against captures of the live kiosk, in front of a .NET 10 API on AWS Lambda served through CloudFront.
Built for 1920×1080. The screen is pixel-matched to the kiosk's fixed full-HD display and runs full screen at that resolution. In any other window it scales itself to fit, letterboxed, so the live demo below looks like the kiosk on a laptop or a phone.
04 / Resume
Filter by skill
· live from /api/experience
– · Remote
AdvancedMD
– · Hybrid
Arise Virtual Solutions
– · Florida
TOPS Software
– · Florida
LIG Marine Managers Insurance
– · Florida
CENTURY 21
Graduated at age 16
Physics · Massachusetts Institute of Technology
05 / Contact
Hiring, contract work, or a question about one of the projects above. Send a note, pick how you would like to hear back, and I will reply, usually within a day.
06 / Colophon
A footnote for the curious. The whole site follows the KISS principle, keep it simple: there is no server behind this page, only static files on a CDN, four Lambdas, and a CloudFront Function.
prefers-color-scheme until a visitor picks one with the toggle beside the edition menu. A few lines of script in the head apply a saved choice before the first paint, and every color follows it through color-scheme and light-dark(). The shared palette borrows from JetBrains Rider; this edition warms it to eggshell through a small file that overrides only the design tokens with light-dark(), so no rule is repeated. prefers-reduced-motion switches off the transitions. Inter and JetBrains Mono are self-hosted as variable WOFF2 subsets and preloaded. Icons are a single inline SVG sprite.www to the apex.X-Frame-Options: DENY, nosniff, and a locked-down Permissions-Policy. The site sets no cookies and runs no analytics.302 to the file in S3. If the notification fails, the visitor still gets the resume./api/experience, with a skill filter and an aggregated /api/skills view. It is C# on .NET 10, run on AWS Lambda (arm64) behind this same CloudFront distribution and reading a single DynamoDB table. The OpenAPI contract was written and committed before any code. Responses carry ETags and cache headers so CloudFront answers most requests itself, errors follow RFC 9457 problem details, and the function URL accepts only requests signed by CloudFront. Write endpoints are specified and stubbed for an admin API to come. The contract is public: OpenAPI spec.<dialog> the kiosk demo uses: main.js creates the iframe when a demo button is pressed and removes it on close, so the explorer and the C# Lambda behind it load only when a visitor asks. The function reads a public sandbox of synthetic records through the Firely SDK, maps the resources to small DTOs, and reports every upstream call so the explorer can show what happened; its write side creates, replaces, patches, and deletes demo patients under a guard tied to the visitor's session. Nothing is stored, and no real patient data is involved.main. It assumes a least-privilege IAM role through OIDC, so no AWS keys are stored in GitHub, runs aws s3 sync --delete in two passes (a week of browser caching for fonts and images, five minutes for pages), and invalidates the CloudFront cache. The third-party actions are pinned to commit hashes.python3 -m http.server before each commit.