TypeScript / React
Production feature work in a live web application, framework upgrades and architecture decisions included.
Software Engineer · TypeScript & Platform Engineering · Machine Learning
Ten years building production web applications in TypeScript and .NET, the platform tooling other engineers ship on, and machine learning models trained on real clinical data.
Also building: Hill Peak Holdings, operator-led acquisitions of durable small businesses.
I’m a Colorado Springs-based remote software engineer with ten years across production web applications, platform engineering, and machine learning. At Planet DDS (Apteryx Imaging) I ship React and TypeScript features in a production imaging platform, and own the architecture calls, framework upgrades, and the production operations that keep it running on Azure App Service.
Most of what I build is infrastructure other engineers work inside: the release automation and internal tooling the engineering org runs on, an internal developer portal on Backstage that scaffolds a fully-wired .NET service, and a custom MCP server that unifies Jira, Confluence, and Azure DevOps behind a single queryable interface.
On the machine learning side, I trained DRUNet denoisers for dental radiographs using a Poisson shot-noise model, the physically correct noise process for photon-counting X-ray imaging. I also built the DICOM de-identification and dataset-provenance pipeline behind them, so there is an auditable record of which data trained which model.
Before Planet DDS: enterprise workflow automation at Fiserv, and mission systems in defense and space at Boeing Defense, Space & Security and Harris Space and Intelligence Systems, where I held a TS/SCI clearance. Before that, four years in the U.S. Army as an infantryman.
Outside of work I’m in the mountains, running a local-first home server, or deep in the kinds of stories that explore systems, resilience, and long arcs. Fallout included.
Education
M.S. Computer Science (2016) · B.S. Computer Science (2015)
Security clearance
Held supporting Boeing and Harris. U.S. citizen; eligible to obtain and maintain.
Military service
2008–2012. Combat deployments to Iraq and Afghanistan; awarded the Combat Infantryman Badge.
Working style
Comfortable owning a system end to end and being the person on call for it.
Planet DDS (Apteryx Imaging)
Ship user-facing features in a production React/TypeScript application, owning architecture and technology decisions, framework upgrades, and production monitoring and remediation across Azure App Service. Designed and shipped a custom MCP server unifying Jira, Confluence, and Azure DevOps into one queryable interface: an agent-facing integration layer over three systems of record, adopted beyond the originating team. Built the automated release pipelines and audit trail now used across multiple engineering teams, plus the data and training pipeline behind our image-quality models.
Fiserv (Account Processing / Enterprise)
Engineered automated C#/.NET Core integrations between HR and workflow platforms, eliminating redundant manual data entry across departments and accelerating delivery timelines. Focused on reliability, automation, and clear operational visibility for enterprise teams.
Boeing Defense, Space & Security
Full-stack Java developer building RESTful microservices for a system modeling space assets, in a process-driven environment that emphasized traceability, CI/CD discipline, documentation, and cross-team coordination.
Harris Corporation (now L3Harris), Space and Intelligence Systems
System sustainment for mission-critical .NET applications: debugged and enhanced C#, Bash, and batch tooling to maintain uptime across legacy Windows and Linux environments, supporting on-site operations.
United States Army
Combat deployments to Iraq and Afghanistan. Awarded the Combat Infantryman Badge for direct participation in ground combat operations.
Production feature work in a live web application, framework upgrades and architecture decisions included.
Internal developer platforms and the paved road other engineers ship on: scaffolding, pipelines, and policy in one place.
Training and evaluating image-restoration models on real clinical data, including the pipeline that feeds them.
Backend services, REST APIs, and background jobs built to run unattended.
Deploying, observing, and remediating in production, including the incident at the far end of it.
Agent-facing integration layers over real systems of record, built to be queried rather than demoed.
Planet DDS
Hackathon-winning internal developer platform built on Backstage. An engineer can scaffold a fully-wired .NET service (repo, Azure DevOps pipeline, branch policies, and security gates) without touching the plumbing. I built the AI-guided onboarding scaffolder, which generates a plain-English Claude Code setup script for non-technical users, and the .NET template’s security gate: PII scanning, automated review, and scanner orchestration wired into the branch-policy pipeline. Auth via Microsoft Entra ID.
Planet DDS
A custom Model Context Protocol server unifying Jira, Confluence, and Azure DevOps into a single queryable interface. Cut context-switching and onboarding time; adopted beyond the team that asked for it.
Planet DDS
Automated pipelines that pull Jira data into a CDN-backed system, plus the release framework and audit trail now used across multiple engineering teams, replacing manual documentation and ad-hoc deployments.
Planet DDS
Trained a DRUNet denoiser for dental radiographs on a fork of KAIR, ETH Zürich’s PyTorch image-restoration toolbox, using a Poisson shot-noise model rather than Gaussian: the physically correct noise process for photon-counting X-ray imaging. Built the dataset loader to read raw DICOM directly, including UID-named files with no extension detected by magic bytes, and to condition the network on a per-image noise-level map.
Planet DDS
A Flask app to run the trained SCUNet and DRUNet denoisers interactively against dental X-rays. Auto-detects model architecture from checkpoint state-dict keys, runs CUDA inference, takes DICOM input directly, and offers classical restoration (Richardson–Lucy deconvolution and guided filtering) alongside the learned models for side-by-side comparison.
Planet DDS
Built the de-identification and anonymization pipeline for model training data, plus automated dataset-provenance reports submitted to compliance: an auditable record of which data trained which model. Started independently, before it was assigned.
Bitway, LLC
An unattended content-generation pipeline running end to end: ingestion, AI tooling, text-to-speech, and rendering. Designed, shipped, and maintained solo.
Independent
Operator-led acquisitions of durable small businesses.
Independent
A home server I run, back up, and monitor myself: self-hosted services, data ownership, and a low-stakes place to keep the operational habits sharp.
I’m an operator-led buyer focused on durable, cash-flowing small businesses with respectful transitions.
Message
Use a short message and I’ll reply by email. If you’re reaching out about a role, add the stack and the problem; if it’s about a business transition, feel free to add context.