Deep research is complex. Data is scattered across systems, insights are buried in text and tables, and most LLM demos stop at shallow Q&A. In production environments, researchers need systems that can discover, connect, and reason across both internal and web-scale data, while maintaining reliability, transparency, and compliance.

This session shares the design and deployment journey of a production-grade multi-agent research assistant built to perform deep exploration and long-form reasoning. Drawing inspiration from emerging deep research systems such as OpenAI’s and Gemini’s, the talk explains how similar capabilities were engineered for enterprise settings where privacy and data governance are essential. Attendees will learn how context engineering (selective retrieval, summarization, context routing, isolation) and reasoning strategies (planning, deliberate thinking steps, specialized agents) work together to manage complex research workflows.

The session builds on the team’s recent research publication in Frontiers in Artificial Intelligence and expands on how these systems reduce discovery cycles from weeks to days.

What You Will Learn

• How to design and deploy multi-agent research systems for deep reasoning and synthesis

• Techniques for managing complex context flows and orchestrating reasoning steps

• Strategies for combining structured and unstructured data sources while maintaining compliance and traceability

Who Should Attend

AI engineers, research technologists, data scientists, and architects building intelligent systems for knowledge discovery, synthesis, and long-form reasoning.