Hardware AutomationCase study
Bobbotics Hardware Test Automation Tool
An end-to-end hardware testing pipeline that triggers server tests, watches the machine's LEDs through a camera, validates the result, and files the evidence on its own.
- Company
- IBM
- Role
- Software Engineer Intern
- Timeline
- May – Aug 2026
Impact
45–90
Minute manual flow automated
CV
Camera-based LED detection
Auto
Jazz ETM evidence reporting
The problem
Validating IBM Power hardware was a manual 45–90 minute routine: kick off ASMI tests, physically watch the server's status LEDs, confirm the outcome, and then write up the evidence by hand.
Because a person had to be in the loop the whole time, runs were slow to schedule, hard to repeat exactly, and produced evidence that wasn't machine-readable.
What I built
- 1
Drive the system like a person would
Used Playwright to operate the ASMI web interface and trigger tests, with Telnet/SSH and REST APIs for direct system control.
- 2
Give the rig eyes
Mounted a camera on a Raspberry Pi and used computer vision to detect server LED states, replacing the person watching the front panel.
- 3
Validate automatically
Checked the observed hardware state against expected results so each run ends with a clear pass or fail.
- 4
File the paperwork
Reported results straight to Jazz ETM, producing machine-readable test evidence with minimal manual intervention.
How it flows
01
Trigger
Playwright starts ASMI tests
02
Observe
Pi camera + CV read server LEDs
03
Validate
Telnet/SSH and REST checks
04
Report
Evidence filed in Jazz ETM
The results
- A 45–90 minute manual workflow became a single automated run.
- Test evidence is now machine-readable and consistent between runs.
- Runs need minimal manual intervention from start to finish.
Stack
- Python
- Playwright
- Raspberry Pi
- Computer Vision
- Telnet/SSH
- REST APIs
- Jazz ETM
Next case study
Continuous Compliance as Code