conda/pip package conflicts for tensorflow-gpu

I am using ubuntu 20.04 in Windows 11's WSL2 environment. The PC has a Nvidia GPU and so I have been using CUDA to accelerate tensorflow code execution. I used conda to install tensorflow 2.6.2 and it got the correct versions of drivers, cudatoolkits, and cudnn installed. I am happy with it.

The issues arised after I used conda to install tensorflow-addons and tensorflow-datasets. The available latest version of tensorflow-addons on conda is 0.13.0. So I conda installed it. However, it is not compatible with tensorflow 2.6.2. According to tensorflow-addons github site, I would need 0.14.0 or 0.15.0. https://github.com/tensorflow/addons

So what options do I have? I could use pip to install tensorflow-addons. It would get me 0.15.0 version of tensorflow-addons. But how do pip packages and conda packages co-exist? If I have different versions of tensorflow-addons installed in conda and pip, which one will be used?