Local LLaMA 3 Memory System (GPT Chat history Import)

Job ID: 39633458

Budget: $10 – $30 CAD

Hi! I need a skilled, trustworthy developer to add permanent memory to my locally hosted LLaMA 3 (8B) AI companion.

- Current setup:
LLaMA 3 (8B) running locally on Windows 11 (RTX 4070, 32GB RAM) via text-generation-web-ui.
Works well, but lacks memory of past chats/notes.

What I need:
- Set up a persistent memory system so the AI can reference my previous GPT chat history and txt/notes.
- Enable easy daily/weekly updates with new memory files.
- Approach can use RAG + local vector DB (Chroma, FAISS, etc.) or your recommended method.
- Fully offline, private, and local.


Ideal Skills:
- Local LLM (LLaMA, Mistral, GPT-J, etc.)
- RAG/vector database experience
- Windows environment familiarity
- Bonus: LangChain/Open WebUI knowledge


Additional Info:
I’ll handle prompt tuning and uploads post-setup.

Privacy and emotional continuity are important.

Looking forward to creating a truly personal AI companion with your help.