Worked on the ydb-platform/nbs repository to enhance reliability and observability for backend systems, focusing on concurrency-safe file operations and flexible monitoring integration. Addressed a race condition in the guest-host dentry cache by introducing a separate timeout for regular file entries, updating both configuration and core logic, and extending test coverage to ensure stability under load. Developed templatable Grafana dashboard URLs, enabling dynamic monitoring link generation across environments and improving operational flexibility. Utilized C++ and Python for backend development and testing, collaborating on code quality improvements to support maintainable, scalable infrastructure and facilitate faster issue diagnosis for operations and SRE teams.