Advanced Avatar Viseme Rigging

I’m looking for help adding full viseme and emotion support to three existing avatar heads that I use in a virtual-reality project built with Unity. Each head already has clean topology and a basic facial rig; what I need now is the advanced layer of blendshapes and setup so the characters speak and emote naturally in-game.

Here’s what I’m after:
• A wide range of visemes that map cleanly to Unity’s lips-sync solutions (at minimum the common phoneme set such as AI, E, O, U, L, W-Q, etc.).
• An equally rich emotion library—joy, sadness, anger, fear, surprise, disgust, plus at least a few subtle variants (e.g., smirk, brow raise).
• Clean, non-exploding mesh deformations with no eye or teeth clipping.
• FBX exports and a simple Unity demo scene showing the visemes driven by an audio clip and the emotions triggered from an Animator or scriptable parameters.

I’ll supply the three head meshes (FBX or OBJ), texture maps, and reference videos. You return:
1. Updated rigged heads with blendshapes.
2. The Unity demo scene and any custom scripts you created.
3. A brief read-me explaining setup and naming conventions.

If you’ve done detailed ARKit or VRChat-style viseme work before, that’s a strong plus. Let me know your expected turnaround and a few samples of similar facial-rigging projects you’ve completed.