Problem

Manufacturing teams analyse ATE test logs and track PCB yield by hand, which is slow and error-prone.

Approach

Parses ATE-style test logs in CSV and JSON, extracts per-board pass/fail records and computes yield KPIs across test steps. Classifies hardware failure modes (opens, shorts, parametric deviations) by board type and test step, generates ranked defect PDF reports with trend charts, and keeps an SQLite traceability store of board ID, test step, result and timestamp.

Results

Analysed logs for 150 boards across three board types. 18 pytest cases, zero regressions.