Set up a public, GitHub-Action-based CI infrastructure

[leofang]

OS/Archs:

### Tasks
- [x] linux-64, 1 GPU
- [ ] linux-aarch64, 1 GPU
- [ ] win-64, 1 GPU
- [ ] linux-64, 2 GPUs
- [ ] linux-aarch64, 2 GPUs
- [ ] win-64, 2 GPUs

Requirements ("public test matrix"):

• Python: 3.9, 3.10, 3.11, 3.12
• CUDA driver version: 11 & 12
• CUDA Toolkit version: 11.8 & 12.3+

[m3vaz]

The linux-64 1xGPU config we should be able to start with, the rest we'll either need to share capacity from other pools or otherwise rack and stack.

[sandeepd-nv]

Please review initial set of changes here: #83 and #84.

[leofang]

The first four tasks are done (single-GPU CI for all supported platforms + multi-GPU CI for linux-64), see #267 (build) & #124 (test). The remaining two tasks are currently blocked by lack of multi-GPU runners for linux-aarch64 / win-64 (cc: @ajschmidt8 @m3vaz for vis). Let's mark this task as done and track the remaining CI improvements separately.