Worked on extending the differentiation capabilities of the Mooncake.jl repository by implementing a backward rule for the BLAS.nrm2 operation. This addition enabled robust reverse-mode automatic differentiation for nrm2, improving the accuracy and flexibility of gradient-based workflows in Julia. The work involved updating BLAS rule definitions, enhancing associated tests, and aligning documentation to support future releases. Focus remained on feature delivery and ensuring release readiness, with no major bugs reported or fixed during this period. Leveraged expertise in Julia programming, linear algebra, and numerical differentiation to strengthen the library’s core functionality and facilitate adoption by downstream users.