Implement Hybrid Search Functionality

## Description
Add hybrid search capabilities to improve our current search functionality by combining dense vector search (semantic similarity) with sparse vector search (keyword matching).

## Why
Hybrid search will provide more accurate and relevant search results by leveraging both:
- Semantic understanding (what the user means)
- Keyword matching (what the user types)

## Acceptance Criteria
- Implement hybrid search using our existing vector database
- Configure appropriate vector models for both dense and sparse embeddings
- Create API endpoint for hybrid search queries
- Support filtering in search results
- Ensure performance meets our response time requirements

## References
- Qdrant hybrid search documentation: https://qdrant.tech/documentation/beginner-tutorials/hybrid-search-fastembed/
```

Start by locating the repository's existing search implementation and vector-database integration, then review the referenced Qdrant hybrid-search documentation. Define the dense and sparse models, API query and filtering behavior, and response-time target before implementing; done means the hybrid endpoint works with filtering and meets the stated performance requirement.

Written by the indexing model from the issue text.