Lothiraldan developed and integrated advanced observability and experiment tracking features across langgenius/dify, Vigtu/langflow, and huggingface/accelerate over a two-month period. In langgenius/dify, they enabled end-to-end tracing for LLM applications by integrating the Opik Tracking Tool, improving monitoring and incident response. For Vigtu/langflow, Lothiraldan implemented Opik Tracer to capture detailed traces and spans, optimizing HTTP calls and enhancing traceability with user and session identifiers. In huggingface/accelerate, they upgraded CometMLTracker to support experiment reuse and offline workflows, ensuring compatibility with older versions. Their work leveraged Python, TypeScript, and robust unit testing to deliver maintainable, production-ready solutions.