From Paper to Production in 30 Minutes: Implementing code-less Gen AI Research

Aarushi Kansal, AI Engineer at AutoGPT, discusses the challenges of moving Generative AI research from papers to production. The talk focuses on how to implement research papers quickly, addressing the hurdles of interdisciplinary knowledge, engineering difficulties, and stakeholder buy-in. It highlights the AutoGPT platform as a solution, enabling engineers to focus on core logic using a 'blocks' system and visualizing agent processes as graphs. The presentation covers implementing concepts like episodic memory and iterative RAG, demonstrating how to build and integrate agentic applications seamlessly into existing systems, and introduces the concept of GenOps.