Fix RAG Service & Voice Interaction Issues in Python Healthcare App -- 2

Fix RAG Service & Voice Interaction Issues in Python-Based Healthcare Application

I need an experienced AI/ML developer to diagnose and resolve specific technical issues in a Python-based healthcare application that uses Retrieval-Augmented Generation (RAG) for content delivery and Vapi.ai for AI voice agent integration.

Specific Problems to Solve:

1. RAG System Issues:
- Vector search returning irrelevant clinical content chunks
- Embedding mismatches causing incorrect template retrieval
- Context window optimization needed for accurate responses

2. Voice Interaction Bugs:
- Speech interruption detection not triggering correctly
- Silence timeout thresholds causing premature disconnections
- Response retry logic failing after user interruptions

3. Content Validation:
- Output filtering not enforcing template structure
- Responses deviating from approved clinical script formats

Required Skills:
- Strong Python debugging capabilities
- Hands-on experience with LangChain, LlamaIndex, or similar RAG frameworks
- Experience with STT/TTS features using Vapi.ai
- Prompt engineering and LLM output structuring

Preferred Experience:
- HIPAA-compliant application development
- Healthcare data systems or EHR integration

What You'll Deliver:
- Root cause analysis document for each bug
- Fixed Python code with inline comments explaining changes
- Test results proving issue resolution
- Setup/deployment notes for future maintenance