T-Spline modeling is unlikely ever be as performant or fluid as in Maya.
While modeling in the T-spline environment "feels" very similar to modeling in a Sub-D modeling environment, mathematically T-Splines are a different beast.
Dub-D models are created by subdividing a mesh using a variety of subdivision algorithms with the most known probably catmull-clark. But even a subdivided mesh is still just a mesh and there are not that many intermediate steps necessary to tesselate that mesh for display in the viewport.
The basis off a T-Spline is a control mesh just as with Sub-D's but then a T-Spline has to be converted either directly into a NURBS or at least some intermediary requiring heavier math before that surface then can be tessellated for display. This poses a much heavier computational burden.
If you are comfortable in Maya I'd continue to model there and then convert the control mesh/cage as a .obj, import it into Fusion 360 and convert it into a T-Spline. I personally use Blender for this and it works perfectly. This allows me to use the full toolset Blender has to offer for Sub-D models. I am sure Maya has a vastly superior tool set for this as well. Heck, if you want parametric T-Splines, just rig the Sub-D using the animation tools and you have full parametric control of the the T-Spline.
This is one reason I've suggested to the Fusion 360 team several times to also use the AnyCAD technology for .obj meshes. Not sure they are getting it!