About
I'm Elliot Peter. I build practical systems: software, websites, tools, workflows, and infrastructure for getting work done.
I've spent 17 years running a small AV production business, and I'm now focused on AI-assisted software and systems work full-time. This page is the longer version of how those pieces connect.
Background
Early life
I started using computers when I was five. My family had Apple computers, and I learned by copying BASIC and LOGO programs out of magazines, typing them in line by line, and seeing what happened.
Movies like WarGames made the whole thing feel bigger than a machine in the house. Computers were puzzles, tools, toys, and portals. I did not have language for "systems" yet, but that was what I was responding to.
By junior high, I was running an Apple pirate BBS off my parents' phone line. That was probably less charming to them than it sounds now. But it taught me something early: computers were not just for working alone. They could connect people, move information, and create little worlds with their own rules.
In high school, my first real internet experiences came through Usenet and telnet-based RPGs. Before the web felt like the web, I was already drawn to the network underneath it: text interfaces, remote systems, communities, handles, commands, and strange little doors into other places.
Career path
My career path has been far from a straight line. I moved through restaurants, manufacturing compliance, brewery operations, web development, and eventually live production.
The roles changed, but the pattern stayed familiar: learn the system, find the pressure points, figure out what matters, and make the work happen.
Work in IT
I have always kept a hand in software and IT work. Sometimes it was a job. Sometimes it was a tool for myself. Sometimes it was just for fun.
My first real IT job was over 20 years ago, building e-commerce sites in Miva Merchant. I wrote PHP and MivaScript, handled database migrations, and learned how much business logic can hide inside what looks like "just a website."
That work never fully went away. I built Drupal sites, general web projects, a Xamarin Forms mobile app for a commercial client, a Wix build, scripts, tools, and whatever else solved the problem in front of me.
Work in live production
In my early 30s, I co-founded an AV production company. PA systems, stage lighting, contract work, subcontract work, outdoor shows, small festivals, bars, venues, municipal events, and everything that comes with keeping a small production business alive.
For 17 years, I handled most of the business side: quoting, invoicing, taxes, insurance, client relationships, scheduling, problem solving, and the occasional emergency that had to be fixed before the doors opened.
Live production is a good teacher. The shows do not happen because nothing goes wrong. They happen because the systems get checked and the problems get found, fixed, or worked around before the audience ever sees them.
The show must go on.
Current work in AI
Around two years ago, I started paying closer attention to AI. At first, I used ChatGPT for ordinary things: pet training plans, meal planning, recipe adjustments, personal projects. Useful, but not yet a shift in how I worked.
In January 2026, my friend Ron sent me a podcast that changed the frame. It showed a project manager using structured AI commands to develop and review software. Not chatting with AI. Assigning work.
That was the turn. I stopped thinking of AI as a better search box and started treating it as an execution layer.
The first serious test was HIITTimer, an interval timer app based on an old Xamarin project. Instead of porting the code, I generated a product spec from the old app and rebuilt it from scratch with Cursor and Codex. It reached internal testing in about two weeks and was ready for wider public testing in just over a month.
The moment that made it real came during testing. I was getting ready to manually click through every option in the app. Then I watched Codex build it, deploy it to an emulator, and start interacting with it on its own.
A task I had assumed would be a manual bottleneck moved into the system.
From there I kept building: a context management server, a workflow system, an orchestrator, and eventually Banzai, the pipeline framework that now runs my AI-assisted development workflow.
I'm based in the Louisville / Southern Indiana area, available remotely, and open to local in-person conversations.
The thread
Restaurants, breweries, manufacturing, e-commerce, live production, software. It looks scattered from the outside, but the pattern is consistent: learn the system, find the leverage, build the process, and own the result.
Every one of those contexts required me to learn quickly, manage real constraints, and deliver work that other people depended on.
The tools have changed. The responsibility has not.
Want to work together?
If you need something built, or you need a better way to build it, reach out. Let's start a conversation.
Get in touch