Topic

Reflections and Practices in Finance in the Era of Large Language Models

Large language models are not only reshaping technological paradigms but also driving profound architectural changes in finance, a sector characterized by stringent compliance requirements, high security standards, and complex business scenarios. This talk will explore the key challenges and pathways to breakthroughs for financial LLMs from three perspectives: technological evolution, engineering implementation, and ecosystem integration. Outline Background and Trends Use Cases Engineering Practices Exploring AI Agents Future Outlook Key Takeaways Understand the key differences and technical safeguards required for financial LLMs in areas such as data security, model compliance, and output accuracy. Gain a framework for identifying and analyzing finance-specific tasks that are well suited to LLM-native capabilities, particularly those involving high knowledge density, complex reasoning, and dynamic interaction. Explore the specialized design and engineering practices of financial LLM plugins and AI Agents, with a focus on permission isolation, risk control, and auditability. Gain forward-looking insights into the architecture, security, interaction, and ethical considerations shaping future finance-native AI systems.