Building Production-Ready AI: Agentic Systems, RAG & Responsible AI

Join PyData Seattle and Microsoft Reactor for a practical session on building reliable, responsible, and production-ready AI systems.
As AI moves from prototypes into real-world applications, engineering teams need to think beyond model capabilities. This event brings together perspectives on both responsible AI adoption and production-grade agentic AI, covering how to design systems that are scalable, trustworthy, and effective in practice.
Ramachandra Nalam, Senior Data Engineer at Google, will discuss best practices and limitations of AI in real-world systems, including reliability, privacy, transparency, and human oversight. Attendees will learn a practical framework for making informed and responsible decisions when adopting AI.
Anamika Kumari, Senior AI/ML Engineer at Microsoft, will explore how to architect and deploy production-grade agentic AI systems using RAG pipelines and multi-step agent workflows. Drawing from experience shipping a RAG-based Copilot that scaled to 90K monthly active users, she will cover retrieval design, agent orchestration, evaluation strategies, and common failure modes when moving from prototype to production.
Whether you're a software engineer, data scientist, ML/AI engineer, researcher, or simply interested in building practical AI systems, this session will provide actionable insights into taking AI from experimentation to responsible production use.