Onyx Edge
Request a consultation

Software engineering, infrastructure, and operations

Most consultants advise. I architect, build, and operate.

One senior engineer who designs, builds, and runs production software and the infrastructure under it, in Go, Rails, TypeScript, and Kubernetes.

  • Professional engineering since 2004
  • One senior engineer, accountable end to end
  • Products that reached millions of users
  • Kubernetes in production since 2017

How I work

Three ways to work with me

Start wherever you are.

01Assess

Know where your software stands

A production readiness assessment of architecture, security, data handling, deployment, and development practices, by someone who runs production systems.

  • Architecture through operations, in one review
  • A plain-English report and executive readout
  • A prioritized roadmap you own, regardless of what follows

02Build

Hands-on engineering, delivered

Services, data models, APIs, interfaces, integrations, and the pipelines that ship them. I write the code, not just the report.

  • Go, Rails, and TypeScript services and APIs
  • React, Next.js, and Swift interfaces
  • One-command development environments with Tilt

03Operate

Keep it running, keep improving it

Kubernetes platforms, infrastructure as code, managed hosting, and SRE support, on AWS, GCP, DigitalOcean, or self-managed clusters.

  • Kubernetes on AWS, GCP, DigitalOcean, self-managed, and k3s
  • Infrastructure as code with Pulumi and Terraform, delivered with Flux
  • Managed hosting, monitoring, and on call for launches
Request a consultation

Not sure which fits? Tell me what you run and I will recommend a starting point.

Luke Heuer, founder of Onyx Edge

Who you work with

I'm Luke Heuer, and Onyx Edge is my practice. I have been a professional software engineer since 2004, with full-time roles in backend, frontend, full-stack, and infrastructure engineering, a web agency I co-founded as its CTO, and independent consulting alongside all of it. Infrastructure has been part of the work from the start, from data centers and servers to infrastructure as code since 2016 and Kubernetes in production since 2017, on AWS, GCP, DigitalOcean, and self-managed clusters. I build and run my own products, including apps on iOS, Android, and the web that reached millions of users and held top-chart positions for more than four years, and the Visiting Objects platform, and now build AI-augmented workflows on top of all of it.

What that means for you: the person who reviews your architecture is the person who writes the report, does the work that follows, and knows the system when something goes wrong. No handoff to a delivery team you haven't met. I build with AI every day, on the same engineering practices I bring to your systems.

22 years
In software engineering
16k+
GitHub contributions in the last year, many through my AI agent delivery system
  • Onyx EdgeFounder and principal engineerActive since 2020
  • Vibe HospitalThe same engineering, packaged for teams whose product was built with AI.Active since 2026
  • Visiting ObjectsA product I built and operateActive since 2022

Software engineering

Engineering for software that people depend on

Architecture and engineering

System design and full-stack delivery

I design and build production software end to end: services, data models, APIs, and the interfaces on top of them. I ship what I architect and keep it running, so the design holds up under real load.

  • System and service architectureGo, Rails, TypeScript, Swift, and PostgreSQL, chosen for the job
  • Full-stack deliveryAPIs, data layers, and interfaces built as one system
  • Reviews and auditsCode, architecture, performance, and security
  • Built to operateI run what I build, so it is designed to be run

What clients bring me

  • BuildA new service, API, or app
  • FixPerformance, database, and reliability problems
  • ReviewCode, architecture, and security
  • Take overInherited and legacy systems
  • ExtendNew capability on a system that already works
  • LaunchLoad testing and on-call for a launch or high-traffic event

How a change reaches production

  1. Commit
  2. Pipeline
  3. Staging
  4. Production

Monitoring and alerting across every environment

The same delivery flow runs Visiting Objects today.

Infrastructure engineering and SRE

Systems that stay up and stay fast

I build the infrastructure underneath production software and keep it reliable: environments defined in code with Pulumi and Terraform, Kubernetes platforms on AWS, GCP, DigitalOcean, and self-managed clusters delivered through GitOps, load testing, observability, and incident response.

  • Infrastructure as codePulumi (in Go) and Terraform, reviewed and tested like application code
  • Kubernetes platforms and deliveryHelm and Flux on AWS, GCP, DigitalOcean, self-managed, and k3s
  • Load testing and event on-callFind the limits before launch, and stay on call through it
  • Observability and incident responseMetrics, logs, traces, and alerting

Managed hosting

The platform your software runs on

Fully managed hosting, with monitoring and incident response. Onyx Edge has provided fully managed hosting since 2020, for more than 15 businesses and sites, and it runs Visiting Objects today.

AI-built software

Your team builds fast with AI. I make it safe to depend on.

AI gives a team speed. It also leaves security gaps, shortcuts that won't last, and oversights. I find them, fix them, and keep the system running.

Systems built with AI

Reviewed, fixed, and ready to rely on

When someone on your team uses AI to build something the business comes to rely on, that is a win worth scaling. I review it the way production will, fix what matters first, and stay on to run it if you want me to.

  • Production readiness assessmentArchitecture through operations, in one review
  • Prioritized roadmapWhat to fix first, with an executive readout
  • Defined projectsProduction-readiness work off the roadmap
  • Ongoing guidanceSenior support as the platform grows

Production readiness assessment

Prioritized roadmap

  • ArchitectureService and data design
  • SecurityAccess, secrets, exposure
  • Data handlingStorage, movement, protection
  • DeploymentEnvironments and pipelines
  • PracticesReview, testing, operability

An AI-augmented delivery workflow

  1. Plan
  2. Build
  3. Test
  4. Review
  5. Record
Every change planned, tested, reviewed, and recorded. It is how I build my own products.

The workflow itself

I build and support the software your team builds with

Speed from AI can be made repeatable. I design, build, and support AI-augmented development workflows: the agents, review gates, tests, and delivery pipelines a team uses to build software.

  • Workflow design and setupFitted to your team, your stack, and your risk
  • Agent tooling and integrationsMCP servers and internal tools your agents can use
  • Review and test gatesSo fast changes are also checked changes
  • Ongoing supportKept current as the tools change

The workflows and MCP servers I build work across Claude, Claude Code, ChatGPT, Codex, and other MCP clients.

  • Claude and Claude Code
  • ChatGPT and Codex
  • Model Context Protocol (MCP)

Record

Production systems I have built, run, and stood behind

More than twenty years of building software and running it in production.

Since 2004
Professional software engineering
4+ years
In App Store and Google Play top charts
~30M
Requests a month at peak, scaled alone
100%
Job success score across 2,000+ contracted hours

AI agent delivery system

AI coding agents ship changes under my review: each one planned, tested first, decided on the record, and promoted through staging.

Onyx Edge managed hosting

A fully managed hosting platform and service that has run more than 15 businesses and sites since 2020, including moving a high-traffic publisher from WordPress to Ghost.

Consumer apps at scale, built and run alone

Apps on iOS, Android, and the web that reached millions of users and held top-chart positions for more than four years, on a back end scaled by one engineer to nearly 30 million requests a month.

Visiting Objects, in production today

A SaaS platform for publishers that Onyx Edge designed, built, and operates end to end: site search, analytics on what readers look for, and AI-readable content.

“A highly skilled and experienced Ruby/Rails web application architect. His expertise in web development, infrastructure, and security is both broad and deep.”
Client review, legacy platform engagement, 2020

The full record

  • 2025 to now

    AI agent delivery system. The system I built to run software delivery with AI coding agents: every change planned, tested, reviewed, and recorded. Read the case study

Show 9 earlier entriesHide earlier entries
  • 2022 to now

    Visiting Objects. Site search, analytics, and AI-readable content for publishers.

  • 2023 to 2025

    Embedded infrastructure and reliability engineer for a software company.

  • 2020 to now

    Onyx Edge managed hosting platform and service. 15+ businesses and sites.

  • 2019 to 2020

    Legacy Rails and PostgreSQL platform. Resolved database performance problems and traced a production deadlock that had gone undiagnosed.

  • 2016 to 2018

    StackPair. Job matching on Go services and language processing, on Kubernetes, tracking openings at 2,000+ companies.

  • 2015

    Lead engineer for a data-driven ad platform. Built the first version of its dynamic creative system and advised its clients.

  • 2009 to 2014

    Top-chart apps on iOS, Android, and the web. Millions of users, 4+ years in the top charts, built and run alone.

  • 2008 to 2009

    Interface engineering at an enterprise knowledge-management company.

  • Since 1999

    Security. Penetration testing, source review, hardening, and disaster recovery.

Contact

Start with a conversation

A clear picture of where your software stands and what to do next.

  1. Tell me what you run

    The software, where it runs, and what is on your mind.
  2. I review and respond

    I reply within a business day with a first read.
  3. You get a scoped next step

    A clear recommendation, no obligation.

Request a consultation