将Conda Python环境部署到AWS Lambda Layer遇Numpy问题求助

我正尝试将Conda管理的Python环境打包为AWS Lambda Layer，供Lambda函数调用。目前用预编译的Pandas/Numpy Pip包部署Layer可正常运行，但用Conda打包环境时出现问题。我倾向于用Conda做版本管理与兼容性校验，不想为AWS额外混用Pip环境。

打包后运行时Numpy出现以下报错：

Error importing numpy: you should not try to import numpy from its source directory; please exit the numpy source tree, and relaunch
UserWarning: mkl-service package failed to import, therefore Intel(R) MKL initialization ensuring its correct out-of-the box operation under condition when Gnu OpenMP had already been loaded by Python process is not assured. Please install mkl-service package

我了解Conda包是预编译版本，而AWS Lambda Layer要求包在AWS机器上编译，这可能是问题根源，但Redis等其他包用Conda打包后可正常运行。恳请有相关经验的人提供解决方案，谢谢！

操作步骤：压缩Conda环境并上传至AWS Lambda Layer

内容的提问来源于stack exchange，提问作者Simon