Muaz Ashraf

I am an AI Engineer with over 3 years of hands-on experience building production-ready Generative AI systems. My expertise spans RAG pipelines, agentic architectures, custom MCP servers, and multimodal AI workflows. I specialize in Python-based backend development, LLM integration, retrieval optimization, and AI automation, consistently delivering measurable performance and productivity gains. My experience includes developing advanced RAG pipelines, improving answer relevance by 80% through hybrid retrieval and cross-encoder reranking. I've designed and implemented agentic AI systems using LangGraph and CrewAI for complex reasoning and tool execution, alongside building custom MCP servers for scalable agent workflows. I also have a strong background in multimodal AI, crafting pipelines for text, image, audio, and video generation, including AI video generation workflows that reduced manual editing time by 70%. Additionally, I've engineered voice-based AI agents for automated tasks and implemented activity detection systems using computer vision techniques like object detection, OpenCV, and Yolo.