Trainer for Arabic Lexicon RAG — Hands-on Training with FAISS/Chroma + LangChain (Arabic Project
Budget: $25 – $50 USD
Project Overview:**
I’m developing the **Arabic Lexicon Project (Al-Qamus Al-Dhaki)** — a structured linguistic knowledge base (CSV) containing lexical cards for Arabic words and roots with definitions, explanations, and Quranic, poetic, and Hadith examples.
I need **hands-on training** to build a working **RAG (Retrieval-Augmented Generation)** system using **FAISS or Chroma** and **LangChain**.
**Training Objective:**
- Step-by-step practical sessions to build a working prototype using my data.
- By the end of the sessions, I should be able to update the data, run retrievals, and generate Arabic answers independently.
**Training Scope (3–5 Sessions):**
1) Environment setup (Google Colab or VS Code) + library installation.
2) Convert CSV data into documents + create embeddings + build FAISS/Chroma vector store.
3) Test similarity search with Arabic queries.
4) Connect retrieval to an LLM using LangChain (Arabic prompt with grounded answers).
5) Deliver documentation + best practices for maintenance and scaling.
**Trainer Requirements:**
- Proven experience in Python, LangChain, FAISS or ChromaDB, and Embeddings (OpenAI or Sentence-Transformers).
- Prior experience with RAG/NLP projects (GitHub links or examples preferred).
- Clear and patient teaching style (focus on **training**, not just code delivery).
- Understanding of Arabic is a plus (reading or testing Arabic outputs).
**Deliverables:**
- Working notebook/repository + minimal documentation (English preferred).
- Initial vector index built from my sample CSV data.
- Example Arabic queries with accurate grounded answers.
**Logistics:**
- Sessions via Zoom or Google Meet (screen sharing).
- Language: English (Arabic understanding optional).
- Timezone: Asia/Riyadh (GMT+3).
- Payment type: Hourly.
- Suggested rate: 15–30 USD/hour, total 10–15 hours (flexible based on progress).
**Screening Questions:**
1) Link to a previous RAG project (GitHub/Notebook) + your role in it?
2) Do you prefer FAISS or Chroma, and why?
3) What embedding model do you recommend for Arabic data?
4) Estimated hours to build a working prototype with Arabic CSV data?
5) Can you offer a 30-minute trial session to review setup and workflow plan?
**Required Skills:**
Python, LangChain, FAISS, ChromaDB, NLP, Embeddings, OpenAI API, LlamaIndex, Machine Learning, Data Engineering
I’m developing the **Arabic Lexicon Project (Al-Qamus Al-Dhaki)** — a structured linguistic knowledge base (CSV) containing lexical cards for Arabic words and roots with definitions, explanations, and Quranic, poetic, and Hadith examples.
I need **hands-on training** to build a working **RAG (Retrieval-Augmented Generation)** system using **FAISS or Chroma** and **LangChain**.
**Training Objective:**
- Step-by-step practical sessions to build a working prototype using my data.
- By the end of the sessions, I should be able to update the data, run retrievals, and generate Arabic answers independently.
**Training Scope (3–5 Sessions):**
1) Environment setup (Google Colab or VS Code) + library installation.
2) Convert CSV data into documents + create embeddings + build FAISS/Chroma vector store.
3) Test similarity search with Arabic queries.
4) Connect retrieval to an LLM using LangChain (Arabic prompt with grounded answers).
5) Deliver documentation + best practices for maintenance and scaling.
**Trainer Requirements:**
- Proven experience in Python, LangChain, FAISS or ChromaDB, and Embeddings (OpenAI or Sentence-Transformers).
- Prior experience with RAG/NLP projects (GitHub links or examples preferred).
- Clear and patient teaching style (focus on **training**, not just code delivery).
- Understanding of Arabic is a plus (reading or testing Arabic outputs).
**Deliverables:**
- Working notebook/repository + minimal documentation (English preferred).
- Initial vector index built from my sample CSV data.
- Example Arabic queries with accurate grounded answers.
**Logistics:**
- Sessions via Zoom or Google Meet (screen sharing).
- Language: English (Arabic understanding optional).
- Timezone: Asia/Riyadh (GMT+3).
- Payment type: Hourly.
- Suggested rate: 15–30 USD/hour, total 10–15 hours (flexible based on progress).
**Screening Questions:**
1) Link to a previous RAG project (GitHub/Notebook) + your role in it?
2) Do you prefer FAISS or Chroma, and why?
3) What embedding model do you recommend for Arabic data?
4) Estimated hours to build a working prototype with Arabic CSV data?
5) Can you offer a 30-minute trial session to review setup and workflow plan?
**Required Skills:**
Python, LangChain, FAISS, ChromaDB, NLP, Embeddings, OpenAI API, LlamaIndex, Machine Learning, Data Engineering