Master thesis: AI-Assisted Nightly Test Result Triage and Fault Analysis

Design and evaluate an AI-assisted system for automated triage of nightly test failures in embedded and automotive software environments. Ingest and analyze test logs, reports, metadata, source code, and historical execution results to classify failures, identify root-cause locations, and produce human-readable explanations. Recommend next actions such as fixing production code, stabilizing flaky tests, or investigating configuration issues. Deliver both a practical prototype implementation and a scientific evaluation of the proposed approach. This thesis explores how ML and LLM methods support engineers in diagnosing complex CI environments, with research questions focusing on accuracy, context selection, robustness, and risk assessment.