The following steps assume the loaded object is a visual instance.
Step 1: Retrieve the mesh's AABB. This will be needed later on.
Step 2: Put your current camera at the center of this AABB:
//bound_box is the retrieved AABB.
cam.translation = bound_box.position + bound_box.size/2.0
Step 3: Move the camera along its local z axis by the AABB's size/2.
cam.translation += cam.get_camera_transform().basis.z * bound_box.size.length()/2.0
If all goes well, you should see the entire mesh from any angle.