There are two steps to get this working, getting the raspicam node up and running then getting the face detection node working.
If I remember rightly the raspicam node is a little unusual in that you have to send a start service request to it before it will publish any images. This makes sense on a Pi so you can conserve processing power when you don't need the images. When you have this node up and running and you can view the image stream on RVIZ on your remote computer then you're ready to move onto step two.
The face_detection node you linked to is designed to work with stereo cameras so it's not suitable for use with just a raspberry pi camera, however the face_detection_tracking package does work with monocular camera images and also produces the same bounding box overlay you're looking for. Running this node on your remote computer will spare the CPU of the Pi and result in a faster detection frequency.
When this node is running there will be a new image topic you can view in RVIZ that will highlight any detected faces.
Hope this gets you going in the right direction.
What are you wanting to see in RVIZ? What type of information do you want to know about the detected face? 2D bounding box? 2D centre? Cropped image?
Can I have both 2D bounding box and 2D centre? Sorry, not very familiar with face detection though. And I have just started with ROS as well.
Also can I make the raspicam_node show entire image stream? I mean continuously like what the face_detector does?