Sambhav worked on expanding tensor operation interoperability in the llvm/torch-mlir repository by enabling TOSA conversions within the Bazel build rules. This feature broadened the set of supported tensor operations, allowing smoother integration with TOSA-backed backends and improving deployment flexibility for downstream production pipelines. Sambhav’s approach involved customizing Bazel build rules and orchestrating the Torch-MLIR workflow to support TOSA integration, leveraging skills in C++, MLIR, and build system configuration. While no bugs were addressed during this period, the work laid a technical foundation for future backend optimizations and reduced friction when integrating Torch-MLIR with TOSA-based technology stacks.