Worked on the apache/pulsar repository to enhance reliability by addressing a critical issue in VIP health monitoring. Focused on backend development using Java, the work involved implementing deadlock detection within the broker’s health probe, allowing the system to identify and report potential service unavailability. By improving the semantics and failure signaling of health checks, the changes enabled faster detection of outages and reduced the risk of silent failures. Demonstrated strong skills in concurrency and system monitoring, with targeted debugging and instrumentation improvements that increased observability and traceability in a key subsystem, ultimately contributing to higher service availability and reduced MTTR.