docs: group the examples/README into a domain table

[YuuGR1337]

Fixes #165

---

[sbryngelson]

Thanks for jumping on this, @YuuGR1337 -- and welcome!

The direction is right, but the issue is really about reorganizing examples/README.md itself -- that's the file with the flat list of 50+ demos that we want grouped by domain (keeping its existing "Start here" hero block at the top). This PR instead adds a table to the top-level README.md and leaves examples/README.md untouched, so the core ask isn't covered yet. The top-README pointer is a nice bonus, but it was the optional part.

Two things to turn this into a merge:

1. Move the grouping into examples/README.md and cover all the demos, not a highlight subset -- including the DSP / embeddings / ONNX group (fft.py, sentence_embeddings.py, rag_embeddings.py, onnx_import.py, spectral_analysis.py, ...) that's currently missing. Each row = filename + a one-line blurb from the script's module docstring. the train_*.py family as a single row is fine.
2. Keep the top-level README table too if you like -- but please drop Fixes #165 unless examples/README.md is the file being changed, so the issue doesn't auto-close early.

Happy to review again once it targets examples/README.md. Thanks again for picking this up!

[YuuGR1337]

Thanks for the review — I've updated the PR to address your feedback. Please take another look when you have a moment.

[sbryngelson]

Thanks for turning this around, @YuuGR1337 -- and for surfacing the ONNX/embeddings demos that genuinely weren't listed (onnx_import, onnx_finetune, sentence_transformers_ane, rag_embeddings). That's the useful part.

Two small things before merge. It turns out examples/README.md is already grouped into domain sections (Pretrained models, Training, LLM inference, Linear algebra, Applied math, Stable Diffusion, ...), each written as a bullet list -- so #165's "flat list" premise was a bit stale, and the goal shrinks to slotting the missing demos into that existing style:

1. Match the bullet style. The new section is a markdown table, but every other section uses - \file.py` - description`. A table stands out; please use bullets to match.
2. Drop the 3 already listed. fft.py (under Linear algebra), spectral_analysis.py (Applied math), and sentence_embeddings.py (Pretrained models) are now listed twice -- keep only the genuinely-new four.

Concretely, replacing the table with this keeps the file consistent and dedup'd:

**ONNX and embedding integrations**
- `sentence_transformers_ane.py` - drop-in `sentence-transformers` `SentenceTransformer` running the encoder on the ANE.
- `rag_embeddings.py` - ANEForge embeddings as a LangChain `Embeddings` adapter for RAG.
- `onnx_import.py` - import an ONNX model and run it on the ANE, validated vs onnxruntime.
- `onnx_finetune.py` - transfer-learning from an imported ONNX model on-engine: frozen extractor + a fresh linear head trained on the ANE.

The top-level README.md overview table is a nice bonus -- happy to keep it. Once examples/README.md uses bullets and drops the 3 duplicates, this is good to merge. Thanks again!