About
Harel Wilner is an AI Engineer at Cywareness, focused on turning emerging AI research into reliable production systems. His work spans agentic AI, LLM evaluation, context engineering, RAG, open-weight models, AI safety, and cloud and on-premise AI deployment. Previously, Harel served as AI Director at Menthoda, where he worked on AI-driven matching, recommendation, and multi-agent systems. Alongside his production work, he conducts independent technical research and open-source experimentation, including a hybrid RAG-CAG question-answering framework and an implementation and extension of SASA for inference-time LLM detoxification. He also writes extensively about the transition from prompt engineering to broader AI engineering, with particular emphasis on evaluation, requirements, context, reliability, and responsible AI deployment."