Senior AI Engineer – Agent Orchestration, Graph RAG

• Build and improve agent workflows for claim analysis, retrieval, verification, and response generation
• Optimize orchestration, including routing, state management, parallel execution, retries, and failure recovery
• Design and optimize graph, SQL, and vector-search queries, indexing strategies, filtering, and reranking
• Build reliable knowledge-ingestion pipelines covering validation, entity resolution, deduplication, incremental updates, and re-indexing
• Preserve source provenance and citation metadata while keeping graph and vector representations consistent
• Reduce response times and cost through profiling, caching, concurrency management, model selection, and efficient context management
• Build evaluations and regression checks for retrieval quality, verdict accuracy, citation correctness, and appropriate handling of insufficient evidence
• Monitor production systems, investigate failures, and collaborate with product engineers and domain specialists

• 6+ years of professional software engineering experience with substantial ownership of production backend or distributed systems
• Experience shipping and operating LLM applications, agent workflows, or RAG systems used by real customers
• Strong Python skills, including asynchronous programming, API development, testing, profiling, and debugging
• Experience with LangGraph, LangSmith or comparable orchestration systems
• Practical knowledge of embeddings, vector search, hybrid retrieval, reranking, and structured model outputs
• Strong SQL skills and experience optimizing database queries
• Experience building ingestion or ETL/ELT pipelines with validation, retries, incremental processing, and recovery
• Demonstrated improvements to production quality, latency, reliability, or cost
• Clear communication and the ability to own technical problems from investigation through production delivery
• Preferred: Neo4j, Cypher, knowledge modeling, or Graph RAG / Graph database
• Preferred: Consumer-facing AI applications
• Preferred: Qdrant, PostgreSQL, Supabase, TypeScript, or Next.js
• Preferred: Multilingual or Arabic retrieval and evaluation
• Preferred: Text, image, audio, or video processing

• 21 paid working days annually covering vacation, personal, and sick leave
• Agreed holidays including major Islamic holidays