The Short Version
I’m an Audio Visual Specialist based in Toronto. I install, configure, and maintain broadcast infrastructure across international studio locations — ATEM switchers, cameras, audio systems, SDI routing, the whole signal chain. Over the past year, I’ve been using AI tools to build web applications that solve real problems my team faces every day. I’m not a software developer. But the tools I’ve built are in production across multiple countries.
Background
I got into video production early — shooting and editing freelance through high school, then studying it at Durham College. After graduating I went straight into live broadcasting at TraderTV.LIVE, where I ran a daily financial broadcast for three years. That’s where I first got deep into Blackmagic hardware — operating ATEM 4 M/E panels and Constellation 8K switchers, building virtual production environments in Unreal Engine, and automating control workflows with Bitfocus Companion.
In 2023 I joined ARRISE powering Pragmatic Play as an AV Specialist. The first three months were spent training in Bucharest, Romania — building studios from the ground up, running cable, installing racks, configuring equipment. Since then I’ve been based in Toronto managing broadcast infrastructure, with deployments to Colombia and Romania for studio buildouts and team training.
The work is hands-on. I’m pulling SDI cable, configuring cameras, keying greenscreens, troubleshooting signal paths, and training international teams on everything from ATEM macros to Ultimatte holdout masks. The department manages hundreds of devices across multiple facilities worldwide.
How the Development Started
In late 2024 I got tired of the Excel spreadsheet we were using to track equipment. Hundreds of devices across the facility — ATEMs, cameras, mixers, hyperdecks — all living in a spreadsheet that was hard to search, impossible to act on, and always slightly out of date.
So I built a simple web app to store it in a proper database. And once every device had a record with an IP address, I realized how many useful things you could build on top of that. What if clicking an ATEM in the list opened a live control panel? What if you could ping every device from the browser? What if the system could schedule image uploads to specific switchers automatically? Each question led to a new feature, and the platform kept growing from there.
I don’t have a computer science degree. I didn’t know Django or WebSockets or Docker before starting. What I did have was a clear understanding of the problems — and a feeling that better solutions should exist — along with access to AI tools that could help me build them.
AI as a Learning Tool, Not a Crutch
There’s a difference between asking an AI to write code for you and using an AI to learn how to build something. I’ve tried to stay on the right side of that line.
When I needed real-time communication between a browser and an ATEM switcher, I didn’t just ask for working code. I learned what WebSockets are, why Django Channels exists, how ASGI differs from WSGI, and what a consumer pattern looks like. When I needed to reverse engineer ATEM protocol commands, I sat in Wireshark analyzing hex dumps — no AI can do that for you. When the content scheduling system needed to span Django, a Go server, and a Windows executable, I had to understand how each piece worked and why it was architected that way.
AI accelerates the process massively. It helps me understand unfamiliar frameworks faster, debug issues I’d otherwise spend hours on, and write code in languages I’m still learning. But the architectural decisions, the domain knowledge, the understanding of what actually needs to be built and why — that comes from years of working with this equipment and these teams. The AI is good at code. It’s not good at knowing what matters in a live broadcast environment.
I think this is where things are heading for a lot of technical professionals. You don’t need to be a developer to build useful software anymore. But you do need to understand your domain deeply enough to know what to build and how it should work. The AI handles syntax. You handle the thinking.
What I’ve Built
This hasn’t been a straight line. The first version of the equipment database was a Flask app — basic, ugly, barely functional. But it worked, and building it taught me how web frameworks actually operate. I eventually found Django and rebuilt everything properly with models, views, templates, authentication. Real-time control pushed me into WebSockets and Django Channels. Deploying across multiple locations meant learning Docker. SSL and routing led to Nginx and Cloudflare. Each problem pulled me into something new.
A year in, the platform includes browser-based ATEM and X32 control, automated content scheduling, overlay management, equipment monitoring, and custom protocol handlers. It’s containerized, deployed internationally, and used daily. I’ve also reverse engineered parts of Blackmagic’s ATEM protocol using Wireshark, built a proof-of-concept connecting Unreal Engine to live broadcast hardware, and contributed to robotic camera tracking through a custom Unreal Engine plugin.
None of this came from a single prompt. It came from months of building, breaking, debugging, and learning — using AI as a tool to move faster, but doing the actual thinking myself.
The projects and journal on this site document all of it — not just what was built, but how and why.
If you want to talk broadcast, AV infrastructure, building tools with AI, or anything else — reach me at [email protected].