LLM Engineer

About the role

Our assistants answer from a customer's own documents, in Arabic and English. As an LLM engineer you own the quality of those answers: prompts, retrieval, evaluation and the guardrails around them.

What you will do

• Build retrieval-augmented generation (RAG) pipelines: chunking, embeddings, search, re-ranking
• Write and test prompts, tool definitions and agent flows
• Create evaluation sets and track accuracy, refusals and hallucination rates over time
• Handle Arabic specifics such as dialects, mixed Arabic-English text and diacritics
• Keep cost and response time under control as usage grows

What we're looking for

• 2+ years in software or ML engineering, with hands-on LLM work (API-based or open-weight)
• Solid Python; experience with vector search and a RAG stack
• A habit of testing: you do not ship a prompt change without measuring it
• Good written English; reading Arabic is a strong plus
• Understanding of data privacy and why sensitive data must stay controlled

Nice to have

• Fine-tuning or model-serving experience
• Voice or speech pipeline experience
• Experience with Arabic LLMs