Build Your Own AI Assistant powered by LinkedIn Content
git clone [email protected]:shahules786/linkedin_ai.git
cd linkedin_ai
pip install --pre-release=allow -e . # mlflow is pre-release
Start by checking out the example notebook to understand how the core functionality works. This will introduce you to:
- How to load LinkedIn post data
- How to initialize the AI assistant
- How to ask questions and get responses
Move on to the experiment notebook where you'll learn:
- How to set up LLM-based evaluation metrics
- How to run and track experiments systematically
- How to compare different experiments
- Ship the best-performing version of your AI assistant
✅ BM25 Search: Uses the BM25 algorithm for fast keyword-based retrieval
✅ Vector Search: Supports semantic search using embeddings
✅ MLFlow Integration: Built-in experiment tracking and logging