QUANTUM Inspector

A desktop tool to watch and inspect a car's touchscreen interface live.

Diagnostic tooling for Android Automotive HMI QA

Professional work · Jan 2026 – present

At a glance

Challenge: UI automation for in-car infotainment kept failing on fragile selectors, and the reasoning behind each fix left no evidence.
Solution: A Python desktop tool for live screen mirroring and UI-hierarchy inspection over ADB, with snapshot evidence and scoped locator generation.
Result: A human-in-the-loop diagnostic layer that improves decisions before they are scaled into Maestro regression runs.

QUANTUM Inspector square-and-circle diagnostic-workflow glyph

The problem

UI automation for automotive HMI fails on selector fragility. Writing robust Maestro flows requires understanding live screen state first, and that reasoning was happening informally, without traceable evidence behind it.

Approach

Where the tool sits: a human-in-the-loop check before decisions are scaled into automated regression runs.

Screen mirroring and live UI-hierarchy inspection against real test-rack devices over ADB.
Snapshot-based evidence capture, so a diagnostic conclusion can be pointed at the state that produced it.
Scoped locator generation to cut selector guesswork before flows are scaled into automated regression runs.

Outcome

Positioned deliberately as a human-in-the-loop diagnostic layer rather than an execution framework, improving the quality of decisions that later get scaled through Maestro automation.

Built with

Python
Qt
ADB
Maestro
Android Automotive