Worked on reliability, observability, and benchmarking improvements across the etcd and bbolt repositories. Enhanced test stability in slackhq/etcd by introducing configurable timeouts and a TimeoutInterface, reducing flakiness in upgrade and downgrade scenarios. In k3s-io/etcd, added a Prometheus gauge metric to track feature enablement, improving production monitoring and feature rollout visibility. For etcd-io/bbolt, expanded the benchmarking tool with new command-line flags for page size and memory mapping, and implemented a sequential deletion write mode to simulate realistic workloads. Leveraged Go, Prometheus, and CLI development skills to deliver targeted backend enhancements that improved test reliability and benchmarking fidelity.