If I contact you about hardware, I already have a real job for it.
SentryOSM is my private home security and automation testbed. Hardware goes into an actual running network with cameras, Frigate, Home Assistant, Sentri, local storage, audio and automation. I am not collecting samples and I am not pretending to be a production OEM.
A real stack, not an isolated bench demo.
The point is interoperability. Equipment has to work beside the rest of the system, recover cleanly, expose useful local interfaces and remain serviceable after the five-minute demo is over.
IP cameras + Frigate
RTSP/ONVIF, multiple streams, detection, recording, clips, snapshots and source stability.
Home Assistant
Device state, controls, notifications, dashboards, MQTT and normal household automation.
Sentri
Event normalization, policy outcomes, authorization state and security-specific context.
PoE + segmented LAN
Link recovery, power behavior, local management and whether devices behave well when isolated.
Linux / Proxmox infrastructure
Self-hosted services, monitoring, backups and repeatable configuration rather than one-off app setup.
Local accelerators + AI
Camera detection and local inference workloads where the hardware actually adds useful capability.
The useful categories are broader than cameras.
If a device can improve the reliability, visibility or local control of the property, it may have a place in the test environment.
Cameras
RTSP, ONVIF, sub/main streams, low-light behavior, event metadata, local recording, audio and talkback.
Networking + PoE
Switches, injectors, managed/segmented network behavior, PoE recovery and link stability.
Door stations + intercom
SIP, ONVIF events, relays, microphones, speakers, talkback and local acknowledgement paths.
Sensors + automation
MQTT, Home Assistant, local APIs, state accuracy, latency and reconnect behavior.
Storage + power
Surveillance storage, UPS equipment, power-loss behavior, service restart and recovery.
Edge compute + client devices
GPUs, inference accelerators, small systems, wall panels and other local clients.
Camera problems get isolated instead of blamed on the first layer that looks broken.
A recent unstable camera case was tested through the browser path, direct RTSP, repeated connection attempts and a physical switch/port-path comparison against a known-good control camera.
Separate player failure from device failure
The browser could sometimes play the camera, but repeated direct-source tests showed the endpoint itself was intermittently refusing or timing out.
- Compare direct RTSP and application playback
- Use a stable control camera
- Swap the network/PoE path
- Track outage windows rather than guessing from one failure
Stop changing the wrong software
Once the instability followed the camera instead of the switch path, the useful conclusion was that the source device/path needed physical attention. That prevented “fixing” a working media gateway to compensate for an unreliable endpoint.
Intercom work is tested end-to-end.
The current home system includes a door-station event path tied into Home Assistant and Sentri, with local acknowledgement/audio work rather than a cloud-only doorbell workflow.

“It powered on” is not a test result.
I care about whether a device integrates cleanly, tells the truth about its state, survives normal failures and remains understandable months later.
Useful engineering feedback, not a guaranteed favorable review.
If a device belongs in the environment, the output should be enough to reproduce the setup and understand the caveats.
Have hardware that fits one of these jobs?
Send the model number or technical documentation. I can tell you where I would use it and what I would test.