During January 2026, Eric Cao developed observability features for the BerriAI/litellm repository, focusing on integrating Prometheus metrics to track total users and teams. He implemented backend instrumentation in Python, enabling real-time visibility into application usage and supporting capacity planning. The solution was designed for production readiness, introducing minimal overhead while providing centralized metrics collection suitable for dashboards and alerting. By adding unit tests and ensuring traceability, Eric improved troubleshooting speed and reliability. His work laid the foundation for data-driven decision making within LiteLLM, demonstrating depth in Prometheus integration, backend development, and code instrumentation for scalable, reliable monitoring infrastructure.