[Bug] Vector index dimensions mismatch with model embeddings

[Jaswanth-006]

Describe the Bug

In app/services/embedding.py, the AI model is configured to use sentence-transformers/paraphrase-albert-small-v2, which generates 768-dimensional embedding vectors.
However, in app/scripts/create_constraints.py, the vector indexes (student_bio_index and faculty_research_index) are configured with 384 dimensions. This mismatch causes any vector search queries or node registrations to fail in Neo4j.

To Reproduce

1. Set up the schema using create_constraints.py.
2. Try registering a user/student or performing a semantic search.
3. Observe Neo4j dimension mismatch error.

Expected Behavior

The vector index dimension should match the model embedding output size (768).