Creative technologist who combines artistic practice, deep
engineering experience, and AI-augmented workflows to take ideas from
concept to polished production — fast. As a photographer, painter, and
engineer, I bring a designer’s eye and an architect’s discipline to
every project. I direct AI tools (Claude, GitHub Copilot, ChatGPT Codex)
as creative collaborators, single-handedly shipping three production web
platforms and a growing set of client sites while accelerating delivery
70-80% — with the organization, documentation, and
infrastructure-as-code rigor to keep everything maintainable across AWS,
Cloudflare, and Google Cloud. Seeking roles at the intersection of
creative vision, technical depth, and AI-driven efficiency.
Core Competencies: AI-augmented delivery (Claude,
GitHub Copilot, ChatGPT Codex) accelerating development 70-80% •
Creative development & UX design • Solution architecture with Python
and AWS serverless systems, Cloudflare edge security, and Google Cloud
APIs • Process efficiency: infrastructure-as-code, CI/CD automation,
documentation & SOPs • End-to-end product delivery with
Next.js/React
Core Competencies:
AI-Augmented Delivery: Direct AI tools (Claude,
ChatGPT Codex, GitHub Copilot) as creative collaborators — tiering model
capability to task complexity (frontier models for architecture and
design decisions, lighter/faster models for high-volume implementation
and automation) — to prototype, build, test, and document 70-80% faster,
while verifying every generated artifact for correctness and
security
Creative Development & UX: Design and build
visually stunning, user-centered applications with focus on intuitive
interactions and responsive design; blend artistic vision with technical
execution
Solution Architecture: Design scalable,
maintainable systems — Python backends, AWS serverless, event-driven
architectures, and clean API design — from concept through production;
extend AWS with Cloudflare at the edge (DNS, WAF, bot protection) and
Google Cloud managed services such as Text-to-Speech, all provisioned as
code
Organization & Process Efficiency: Turn chaotic
workflows into repeatable systems through infrastructure-as-code, CI/CD
automation, standard operating procedures, and documentation that teams
actually use
Artist & Problem Solver: Apply creative
problem-solving from art practice to technical challenges; user empathy
drives architecture decisions; storytelling through data visualization
and interface design
CASE STUDIES
Longer write-ups on the problems behind the platforms I’ve built —
not just the tech stack, but why they exist.
AI-Augmented Workflows: Claude AI API, ChatGPT
Codex, GitHub Copilot, Prompt Engineering, AI Agent Direction &
Verification, Model Selection & Cost/Latency Tiering, Claude
Scheduled Tasks, MCP Servers (Stripe, Google Ads, Merchant Center,
Search Console), Rapid Prototyping
Data, Content & Tools: DynamoDB (Streams, GSIs,
TTL), MySQL, NoSQL Schema Design, API Integration (Stripe, PayPal,
Google Ads, Merchant Center, Salesforce, FedEx), Sveltia CMS, Git,
Linux, Windows Server, Bash, PowerShell
PROFESSIONAL EXPERIENCE
AssetWatch | Westerville, OH
Lead Backend Engineer | July 2023 – Present
Lead backend development for an IoT monitoring platform, architecting
scalable Python-based systems for real-time data.
Architect and develop Python backend services for real-time IoT data
processing, building Lambda functions and ECS Fargate microservices that
handle high-throughput event streams from AWS IoT Core and Kinesis
Re-architected data flow and system architecture to improve
application responsiveness by 60%, optimizing event processing pipelines
and implementing efficient caching strategies
Build RESTful APIs using Python and FastAPI with API Gateway,
enabling communication between IoT devices and client applications
Implemented zero-downtime canary deployments with rollback,
gradually shifting production traffic while monitoring CloudWatch
error-rate and latency thresholds
Leverage AI coding agents (Claude, ChatGPT Codex, GitHub Copilot)
throughout the development lifecycle — from architectural design through
code generation, testing, and documentation — accelerating delivery by
70-80% while verifying all output for correctness and security
Brought 95% of the environment into infrastructure-as-code (IaC)
using Terraform and pipeline automation tools, implementing fully
functional non-prod environment that reduced production errors by
80%
Write technical documentation with visual diagrams that make complex
systems understandable for technical and non-technical audiences
Tracy Elsinger
Development | Dublin, OH (Freelance)
Creative Technologist & Independent Developer |
April 2020 – Present
Design, build, host, and maintain production websites for independent
clients and my own platforms, from design through secure cloud hosting,
with clients able to update their own content.
Client sites: Built four sites for independent
clients — real estate and fine-art photography
portfolios, a personal and
résumé site, and a private job-search résumé — each with a
Turnstile-protected contact form and Sveltia CMS, so non-technical
clients publish their own changes through a reviewed workflow without
touching code or being able to break the live site
Privacy built in: Build checks block client photos
that carry GPS location or personal names, and fail the deploy if
contact details leak into a public résumé; share/print-to-PDF résumés
keep phone and email off the page
Shared cloud platform: Run all eight sites on one
consistent platform of AWS behind Cloudflare (DNS, DNSSEC, WAF, Bot
Fight Mode, Turnstile, Web Analytics), defined in Terraform, with
keyless CI/CD, shared alerting, budgets, cost-anomaly alerts, and audit
logging
Cost awareness: Built an automated weekly AWS cost
email that breaks spend down by project and compares it with the prior
week. It showed alarms were about half the AWS bill, which led me to
consolidate redundant alarms across every site; I also replaced about
$16/month of AWS WAF with Cloudflare’s free WAF
Ongoing security reviews: Set up a recurring Claude
scheduled task that does a deep-dive security and documentation review
of every site, checks each finding against the live systems, tracks it
in a shared ledger, and implements the fixes — so security keeps pace
with new features instead of waiting for a one-off audit
Led site reliability engineering for 600+ AWS and GCP instances,
bringing systematic problem-solving and user-focused thinking to
infrastructure and compliance challenges.
Designed and developed centralized account management application
(Python/Flask) with intuitive interface for vendor access provisioning
and termination, reducing access review time by 70% and ensuring
consistency in execution across all external accounts
Standardized processes and security on Windows and Linux systems
using Chef and Ansible automation and reporting tools, maintaining
HITRUST and SOC2 compliance requirements across 600+ instances
Collaborated with project managers, application SMEs, and
infrastructure teams to develop standardized project designs.
Maintained project schedules and budgets through effective technical
planning and risk mitigation
Improved solution architecture processes, reducing project kickoff
time by 40%
Gap, Inc | Groveport, OH
Server Engineer | Dec 2013 – June 2015
Server engineer responsible for VMware, Windows, and Linux
infrastructure management and development.
Continuously improved Windows and VMware environments with automated
reporting, security hardening, monitoring, and alerting
Created efficiency reporting dashboards that allowed management to
make better hiring and purchasing decisions based on actual environment
utilization data
Technologies: VMware vSphere, Windows Server, Linux,
PowerShell, Bash
Cardinal Health | Dublin, OH
Server Engineer | July 2008 – Dec 2013
Managed virtual environment consisting of VMware vSphere (450 hosts
and 4300+ guests) and 10 vCenter servers across multiple remote
sites.
Managed the VMware environment with security patching, upgrades, and
large-scale migrations
Participated in projects for planning and deployments in the virtual
environments
Researched and tested new technologies to continually improve the
environment
Technologies: VMware vSphere, Windows Server,
PowerShell
Full-stack recipe, meal-planning, and food-journal platform with
Claude AI built in, deployed on AWS entirely from
infrastructure-as-code.
Rebuilt the food sensitivity tracker into a personalized
pattern-detection engine: statistics on the user’s own history
(repeat-meal consistency, stress and sleep effects, cumulative meal
load) feed an AI layer that proposes testable hypotheses with supporting
and contradicting evidence, and scores diet approaches such as
low-FODMAP against the user’s actual results
Built AI features across the site: a recipe ingredient parser
matched to a cleaned USDA nutrition catalog, recipe modification for
dietary restrictions, and a skin sensitivity tracker that reads cosmetic
labels from a photo and compares reaction vs. safe products to find
likely triggers
Added barcode and label scanning with a shared saved-products
library, household sharing of meal plans and products, and a Free / Base
/ Premium plan system with plan-gated features
Built Stripe subscription billing (tested in a
Stripe sandbox, switched off until launch): hosted Checkout and Customer
Portal for single and household plans, monthly or annual, with prorated
upgrades and Stripe Tax; a signature-verified, duplicate-safe webhook
re-reads each subscription from Stripe, so events that arrive late or
out of order can’t give someone the wrong plan. The Stripe products,
prices, and webhook are defined in Terraform, and the secret key stays
in GitHub and AWS Secrets Manager
Used the Stripe MCP server with Claude to check the
design against Stripe’s implementation planner and best practices before
building, then to inspect the sandbox setup read-only; added a read-only
script that compares the Stripe account with the Terraform config before
launch, and sandbox test checkouts caught a cancellation bug that was
fixed before launch
Designed per-user AI budget caps and a usage dashboard across every
AI feature, so no single user can exhaust shared API spend
Covered sign-in, recipes, and meal planning with Playwright tests
across Chrome, Firefox, and Safari in CI; protected sign-up and contact
forms with Cloudflare Turnstile, WAF, and Bot Fight Mode, with
cross-region backups and alerting
Memory Lab — Language Learning Platform | Personal Project
(Production)
2026 – Present | memory.tracyelsinger.com
Spaced-repetition language platform for Italian, Japanese, German,
French, and Greek, with AI coaching and natural-sounding voices, running
on AWS and Google Cloud.
Built the spaced-repetition engine (SM-2, in Python) that schedules
7+ exercise types — vocabulary, kanji stroke-order writing, sentence
rebuilding, verb conjugation, and more — plus a Daily Mix across
languages and six cognitive training games
Expanded beyond AWS into Google Cloud: added Google
Cloud Text-to-Speech voices for all five languages (WaveNet and Chirp 3
HD), with caching, pre-loading for lag-free playback, per-user quotas,
and a device-voice fallback that keep cost near zero; the Google project
is managed in Terraform with keyless CI, API and quota limits, budgets,
and alerts
Shipped an AI-graded writing exercise that tiers models by task — a
light model writes prompts, a stronger one gives structured,
level-appropriate feedback — with a dedicated privacy consent flow and
self-service deletion for user stories
Added AI explanations and AI-suggested memory hooks with per-user
token budgets and an admin cost dashboard, plus a native-speaker
reviewer workflow where AI correction suggestions wait for human
approval
Added Greek as a fifth language with an alphabet-by-sound module,
and reopened sign-up behind Cloudflare Turnstile and admin approval
Art E-Commerce Platform | Personal Project
(Production)
2022 – Present | tracyelsinger.com
E-commerce site for selling my original artwork and photography, with
payments, digital downloads, and interactive discovery.
Built the full store: PayPal and Apple Pay checkout, persistent
cart, FedEx-based shipping by size and destination, and secure
time-limited digital downloads with automated email delivery
Added safeguards so an original that has already sold can’t be
fulfilled twice — the order raises an alert for follow-up instead
Built interactive features: a before/after slider comparing
reference photos with finished paintings, a color-palette generator, a
travel map of photo locations, and fuzzy search across 500+ works
Rebranded and moved the site to tracyelsinger.com with redirects and
SEO fixes, refreshed the design and blog, and put it behind Cloudflare
(DNS, DNSSEC, WAF, Email Routing) with a Turnstile-protected contact
form
Google marketing, driven through Claude: set up
Google Search Console (verified domain and sitemap), a Google Merchant
Center product feed generated from the site’s catalog (635 works), and a
Google Ads Performance Max campaign; connected Claude Code to all three
through read-only MCP servers, each signed in with only its own narrow
permission, on a Google Cloud project managed in Terraform with keyless
CI and no billing account
Conversion tracking: added Google tag events for
page views, product views, add-to-cart, and purchases, matched to
Merchant Center products and counted once per order, with no tracking on
test builds or buyers’ download links. Auditing the setup through the
MCP servers found the site’s security policy was silently blocking every
conversion from reaching Google Ads; fixed and verified with a headless
browser test
Technologies: Next.js, React, TypeScript | AWS
(Lambda, DynamoDB, S3, CloudFront, Cognito, SNS, SES) | Cloudflare |
Terraform (AWS + Google), GitHub Actions | PayPal, Apple Pay, FedEx API
| Google Ads, Merchant Center, Search Console, Tag Manager, MCP servers
| Leaflet.js, Fuse.js | Python
EDUCATION
The Ohio State University | Columbus, OH | Bachelor
of Science in Business Administration, Information Systems
VOLUNTEER WORK
Volunteer for the Dublin Area Art League with website updates and
technical support