Mechatronics Engineering student at Universiti Teknologi Malaysia, working on applied AI for high-stakes systems — where a wrong answer has a cost, and the model has to explain itself.
Research interests: LLM reasoning and test-time compute, agentic systems, industrial fault diagnosis, and clinical decision support.
Test-Time Compute and Semantic Safety Filtering for Industrial Fault Attribution Accepted — IEEE IRAI 2026
An LLM diagnostic pipeline for industrial machinery that reasons over multiple fault hypotheses before answering, then passes every recommendation through a rule-based safety shield. Evaluated on the CWRU bearing dataset. Paper · Code
