Refactor kernel.py to implement GEMM Triton kernel for FlashInfer competition

Update solution name and definition in config.toml for gemm-triton-v1

Update entry point in config.toml to specify kernel.py::run

Update config.toml with DSA TopK Indexer details and entry point adjustment

Implement DSA TopK Indexer with FP8 quantization and dequantization logic

Initial commit