Software engineering, testing & applied AI
I build software, test the systems it depends on, and apply AI where I can measure the result. My work spans medical devices, full-stack applications, and native iOS.
Life Wrapped began as a way to remember my own work and thoughts. It became a native app for anyone who thinks better out loud.
Talk through your day.
Keep the parts that matter.
Voice recording, on-device transcription, AI summaries, and a Year Wrap. Audio stays on device. Optional cloud summaries send text to your chosen provider.
Published products, professional experience, and public experiments. Each case study explains the work and its boundaries.
28 projects
Published iOS app
A private voice journal that turns spoken thoughts into useful summaries, insights, and a story of your year.
Public engineering lab
A synthetic storefront and release gate that use fault injection, browser tests, and repeated AI evaluations to explain what can ship.
Public engineering lab
A reproducible test environment for a sample task app, with browser and native mobile checks, Kubernetes recovery, and a review gate for AI test drafts.
Public iOS app code
A SwiftUI wardrobe and outfit planner with on-device garment cutouts, suitcase and laundry workflows, and tests that exercise the app on iPhone and iPad.
Public code · in development
A full-stack platform for organizing software projects and generating repository documentation with AI.
Computer vision experiment
A local camera detection experiment combining YOLOv8, OpenCV, and a Flask interface.
Private backend project
The backend for a native wardrobe app, with a versioned API contract, authenticated device requests, and safeguards around paid image operations.
Public macOS developer tool
A local dashboard that turns Codex and Claude Code activity into a pixel office and an inspectable session history.
Reference system with a public guide
A spec-driven reference app where teams sketch an architecture on a shared canvas, ask a local model to extend it, and generate a Markdown spec.
Private developer workflow
My factory playbook as an installable kit: a feature moves through specification, build, four reviews and an independent approver before it reaches a development branch.
Public prototype
A Python tool for parsing and normalizing POCT1 specifications, with a CLI for device onboarding and versioned updates.
Private AI evaluation prototype
A synthetic ad ranking and creative review lab that compares a typed model with heuristic and oracle baselines.
Public local imaging tool
A browser editor and terminal tool for local image generation and reference-guided edits through ComfyUI.
Private workflow collection
A collection of reusable n8n demos for business intake, quoting, follow-up, and reconciliation with human review before outbound action.
Private security lab
A reproducible container security lab that builds, scans, inventories, and signs a small Go service.
Public prototype
Turns a GitHub codebase into chapter-based tutorials, relationship diagrams, and approachable explanations.
Public focus timer
A deadline-driven focus timer that divides remaining time across weighted tasks and recovers a guest session after reload.
Public media automation prototype
A pipeline for producing synthetic character footage with local generation, visual similarity checks, and a human approval queue.
Public creative tooling
A local production workflow linking narration, image generation, review, composition, and editable video assemblies.
Public integration utility
Turns generated video projects into editable Final Cut Pro timelines using Apple’s FCPXML format.
Public tower defense learning game
A mobile-friendly tower defense game where you write real Python in the browser to unlock towers and waves.
Public personal calculator
A local-first calculator that compares home espresso costs with a coffee-shop baseline and projects when a machine pays for itself.
Private data migration tool
A tool for previewing and transforming Follow Up Boss contacts into Sierra CRM import files.
Professional work · Abbott
Device simulation, protocol clients, API tests, and utilities for dependable point-of-care software integration.
VCU team capstone · 2017
A self-powered UAV sensor payload combining spectroscopy, laser ranging, and orientation data for environmental mapping.
Academic firmware projects
Three STM32 builds that turn distance sensing, reflectance sensing, and timed motor control into robot behavior.
Academic terminal simulation
A C terminal simulation that loads seed patterns from files and lets you step through generations.
Delivered community website
A community website built around donations and content that nontechnical volunteers could maintain.
The public repository collection also includes course work, early experiments, and forks. Browse the full GitHub profile
The foundations behind the current work: low-level systems, databases, web frameworks, and small tools. These are historical or learning projects, not current production products.
Academic project · C++
Byte-by-byte state-machine parsing of GPS logs and serial input.
Learning project · Angular / Firebase
Authenticated client management with Firestore-backed CRUD.
Learning project · MEAN stack
A CRUD application using Node.js, Express, MongoDB, and AngularJS.
Learning project · C# / SQL
A console app exploring .NET and a database hosted on AWS RDS.
API learning project · Web
A web application exploring data retrieval with the PokéAPI.
Personal web project
A recipe collection built for easy recollection.
Every discipline here connects to something I have built, studied, or worked on.
Frontend interfaces, backend services, relational data, authentication, API integrations, and product workflows.
Next.js · TypeScript · PostgreSQL · C# / .NET
Device simulation, protocol validation, automated browser tests, and replayable evidence for release decisions.
Playwright · xUnit · Python · XCTest
On-device inference, retrieval, OCR, image segmentation, generation pipelines, and model evaluation against baselines.
Swift / MLX · Vision · Python · FAISS · YOLOv8
SwiftUI apps with local storage, photo and voice workflows, reusable Swift packages, and repeatable UI tests.
Swift · SwiftUI · Swift Package Manager · SQLite · XCTest
Build and test pipelines, reproducible container environments, and recovery checks on a local Kubernetes cluster.
GitHub Actions · Docker Compose · Kubernetes · Terraform
Container hardening and vulnerability scanning, with software bills of materials and offline verification of signed artifacts.
Trivy · CycloneDX · Cosign · Docker
Webhook intake, schema mapping, data cleanup, and review steps before sending model-generated drafts.
n8n · Python · Pandas · Webhooks · Ollama
Turning complex processes into understandable products, useful documentation, and editable media workflows.
Technical writing · Video automation · FCPXML · Web design
A computer engineering foundation spanning sensors, serial data, instrumentation, robotics, and low-level programming.
C · C++ · Assembly · Raspberry Pi · STM32
I build with AI, run models on my own hardware, and design the checks that make agent work reviewable.
APPLIED AI PRACTICE
I began using ChatGPT around its public launch in 2022 and adopted Codex early as coding agents became available. I now choose between chat assistants, coding agents, and local models based on the task, the data involved, and how I can verify the result.
Research, problem framing, and working through ideas before turning them into specs.
Repository audits, code changes, test runs, and delivery with reviewable evidence.
Codebase analysis, implementation plans, and reusable project and iOS playbooks.
In-editor completions and suggestions while working directly with the code.
Another model to consult when researching an idea or comparing answers.
A personal agent I explore for research and tasks that span tools.
Local model serving for repeatable inference, evaluation, and workflow experiments.
A desktop environment for loading local models and trying their settings and output.
My local model library includes Llama, Qwen, and Moondream. I select models for the task and hardware, then check whether their outputs are accurate and useful on examples I can repeat.
I set up the context, skills, tool access, and checks around a model. In Release Evidence Lab, I compared model triage with rules and kept the final release verdict tied to reproducible checks.
I use my notes and task skills to bring decisions and standards into a new session. For longer work, I break the job into phases and leave a handoff with the changes, verification, and next step.
Skills · Vault / Obsidian Vault · Add Task · Start Task · Close Task · Phased Sessions
I bring my design standards into the brief, audit the existing interface, and check the implementation at narrow and wide widths. That gives the agent a reference for decisions about layout, accessibility, and interaction.
Skills · Vault / Obsidian Vault · Impeccable · Design Taste Frontend
My Claude playbooks map Swift changes to the affected tests, check real screens in the simulator, and review privacy, entitlements, and release settings. I use those checks to decide what is ready for a build or needs another pass.
Skills · iOS Feature · iOS Regression · iOS Visual QA · iOS Ship Check
I narrow large code searches to relevant files and check findings against the source. My factory playbook gives specification, implementation, security, UX, and code review separate roles so each step leaves evidence for the next.
Skills · Jev Filter · Factory · Phased Sessions
I turn useful videos into searchable notes. The workflow transcribes the source locally and checks what I already know before adding new material, so future work can build on it.
Skills · Learn From Video · Vault / Obsidian Vault
I use local ComfyUI for drafts and prompt experiments. I can compare outputs on my own hardware and reserve external generation for work that needs a specific capability.
Skills · Local Image Gen
I turn a brief into a storyboard and a written visual direction, build a renderable composition, then review draft frames before export. I keep the timeline editable when a project needs a human finishing pass.
Skills · HyperFrames · Motion Director · Media Use
PERSONAL KNOWLEDGE SYSTEM
My Obsidian vault is an extension of how I think and work. It holds years of research, engineering standards, decisions, and lessons I want to be able to use again.
I record research, design and coding standards, debugging lessons, architecture decisions, and release checklists while the details are still fresh.
I link notes to the projects and topics they inform. When a problem comes back, I can find what I tried, what worked, and why I chose that approach.
I give Claude Code or Codex the relevant standards and decisions as context. After testing the result, I keep the lesson that will help on the next job.
My professional foundation is software reliability in healthcare. My independent work extends that foundation into complete products.
Integration testing, failure investigation, automation, and tools supporting point-of-care software reliability.
Device and driver testing, firmware compatibility, protocol workflows, and reusable test utilities.
The foundation of my work in connected medical devices and healthcare software testing.
Personal products and tools across web, mobile, backend services, AI workflows, and creative automation.
Computer engineering, software testing, machine learning, and continued hands-on exploration.
2013 — 2017
Virginia Commonwealth University
Systems, programming, electronics, networks, and a team capstone in aerial spectral mapping.
View the capstone publication2010 — 2013
Piedmont Virginia Community College
Computer science and technical foundations, followed by a Computer and Network Support Technologies credential in August 2014.
Education on LinkedInISTQB / ASTQB · October 2019
DeepLearning.AI / Stanford · February 2025
Credential ID: J6J1DX3O0OHG
Udemy · 2nd edition · January 2018
Credential ID: UC-J9KQK7SY
2017 · Listed in my LinkedIn honors. Team engineering recognition.
Honors on LinkedInLake Monticello Community Foundation · Recognition for building its website.
Foundation recognitionPreviously held · Recorded validity February 2025–February 2026. Presented as past learning, not a current active credential.
2009 · Early team experience with architectural design and 3D modeling, listed on LinkedIn.
Honors on LinkedInJOSÉ U. RAMIREZ-VILLA
CHARLOTTESVILLE, VIRGINIA
I like turning an idea into something people can actually use.
My background connects computer engineering and medical device software testing with a growing practice in full-stack development, native iOS, and applied AI. I enjoy moving between disciplines—and bringing what I learn in one back into the others.
Life Wrapped came from a personal need. My community website work came from listening to people who needed a simpler way to maintain content. My tools often begin with the same question: what would make this easier, clearer, or more reliable?
I also enjoy the creative side of technology: interface design, visual storytelling, 3D modeling, and video workflows. The engineering matters, but so does how the result feels to the person using it.
TECHNOLOGY IN SERVICE OF PEOPLE
I donated 100+ hours to the Lake Monticello Community Foundation website, with an emphasis on donations and maintainability for volunteers.
Read the Foundation’s storyMy direction is toward AI and machine learning engineering, while keeping the testing discipline and systems understanding that shaped my work.
Native experiences and local-first AI that solve ordinary problems without unnecessary complexity.
Exploring AI agents, MCP, and reusable foundations that make a small engineering practice more capable.
Better retrieval, clearer evaluation, and AI workflows whose outputs can be inspected and improved.
Let’s talk about software, automation, applied AI, or a useful idea. LinkedIn is the best place to start a conversation.
View my résumé
