THE ENGINEER BEHIND DRABSNARG

Understand the system. Then build on it.

I’m CJ Dreher, a media systems engineer in Brooklyn. I work on broadcast networks, audio over IP, and the systems that keep live TV on air.

What I do

Since October 2022 I’ve worked in broadcast at a New York media group, where I’m one of two engineers covering the broadcast plant. I own the broadcast media networks and the Dante audio-over-IP plant, I’m the primary technical owner for Harmonic VOS/XOS playout, and I work as Engineer in Charge during live events. I also take Tier 2/3 escalations.

I like KISS troubleshooting, controlled testing, and writing things down so the next person doesn’t have to guess.

Outside work I build my own tools: AI workflows, voice assistants, and scripts that save me time. Drabsnarg is where all of it gets written down.

Working across the connections

Media systems bring together networks, software, hardware, timing, storage, and people. A dropped audio source can be a clock problem, a switch problem, a Windows firewall profile, or a firmware update from last night. Supporting that kind of plant means following the whole path, not just the device in front of you.

Some recent examples: taking a multi-facility Dante plant off link-local addressing, keeping the channel on air through a master control move, and building a cathedral’s network in one overnight window.

HOW I WORK

Six rules I keep coming back to.

01

KISS first.

Start with the simple explanation. Go deeper only when the evidence says to.

02

One change at a time.

Every change gets a rollback plan, and its own result.

03

Known, suspected, unconfirmed.

Keep them separate in notes, in tickets, and in conversation.

04

If it isn’t written down, it isn’t fixed.

The next person shouldn’t have to guess. Neither should I, six months later.

05

Prevention beats reaction.

Automate the repeat work so the time goes to the new problems.

06

Visibility first.

Logs, monitoring, and traceability before cleverness.

SKILLS / WHAT I WORK WITH

The toolkit behind the work.

Grouped by discipline. Write-ups for most of these are in Projects and the notebook.

Media & broadcast networking

Cisco Nexus (NX-OS)Catalyst (IOS-XE)vPCHSRPIGMP snooping / querierVLANsPort-channelsQoS (planning)Meraki MX / MSSD-WANDual 5G WAN failover

Audio over IP

DanteDante Domain ManagerPTP leader clock designDante Virtual SoundcardSubnet and addressing designPlant-wide migrationYamaha QL5Rio stage boxes

Transport, encoding & playout

SRTRTMPNDI / NDI HXLiveUOpenGearHarmonic Spectrum / VOS / XOSAWS Elemental LiveCrispin RPX / RapidPlayNexio MOSTelestream Vantage

Routing, sync & signal flow

Evertz EQX10VistaLinkMagnumQuartz CPEvertz MSPG / 7001Black burst and tri-level syncTektronix waveform monitors

Live production

Engineer in ChargeRTS OdinSony RCP / CCURoss CarboniteAcuityXPressionInceptionOverDriveVoyagerAvocent HMX KVM

Endpoints & cloud

Intune (Win32 packaging, detection rules, assignment)Microsoft 365SharePointEntra IDWingetGroup PolicyWindowsmacOSLinuxHyper-VDebianrsyslogSamba / CIFS

Media asset management

iconikISGSynologySony CiAspera

Scripting

Python (Netmiko, REST APIs)PowerShellBashNode.js (Alexa / Lambda)

Creative

Premiere ProAfter EffectsMedia EncoderPhotoshopIllustratorPro Tools
ENGINEERING EVOLUTION

One system leads to another.

A thematic timeline, not a dated employment history. The disciplines overlap and continue to develop.

01 / FOUNDATIONS

Hands-on operations

Equipment logistics, scheduling, fault reports, customer coordination, and operational ownership.

02 / NETWORKS & IT

Understanding the paths

Switching, routing, VLANs, endpoint troubleshooting, and enterprise systems administration.

03 / BROADCAST SYSTEMS

Where timing matters

Production systems, live streams, NDI, Dante, SDI, XPression, and the infrastructure behind them.

04 / AUTOMATION & SOFTWARE

Turning fixes into tools

Python, PowerShell, repeatable tests, release procedures, and operational documentation.

05 / LOCAL AI & EXPERIMENTS

Owning the stack

Open-weight models, local inference, self-hosted services, and experiments that remain in progress.