laura shamykhanova

I am an Applied AI / LLM engineer specializing in building production AI agents and LLM-powered systems. I design RAG pipelines, integrate multiple AI providers, and build end-to-end AI workflows using Python and modern LLM APIs, with a proven focus on improving model performance and reducing infrastructure costs (~49%). I translate business problems into scalable AI solutions with a strong product-oriented mindset that blends experimentation, system design, and real-world deployment of generative AI.

At KAZDEVLAB, I built and optimized AI assistants with Retrieval-Augmented Generation (RAG) using the Python and OpenAI API, improving accuracy and contextual relevance by up to 30% in internal tests. I designed task-oriented AI agents for multi-step reasoning, tool usage, contextual decision-making, and modular architecture (prompt orchestration, message routing, context management, and failure recovery), while increasing successful dialogue completion by 40% and reducing failures. I also monitor production AI behavior with evaluation pipelines, run A/B experiments on real users, and collaborate with product and analytics teams to track AI impact on key business metrics and drive product improvements.