Agentic AI Systems: Performance Engineering for Conversational Workloads

This talk was delivered live at the ClickHouse Bangkok Meetup on 19 March at Common Ground, Bangkok by Siddhant Agarwal, Senior Developer Relations Advocate, APAC @ ClickHouse

The full agenda can be found at https://www.meetup.com/clickhouse-thailand-meetup-group/events/313586183/

Talk Title: Agentic AI Systems: Performance Engineering for Conversational Workloads

Talk Description: Generative AI got the buzz, but agentic AI is changing how we actually work. Instead of waiting for prompts, agents explore data, run iterative queries, and reason step by step. That shift creates real infrastructure challenges, including unpredictable query patterns, sub-second response requirements, mixed read and write workloads, and the need for visibility and trust in AI-driven results.

In this session, Sid shared practical lessons from running agentic workloads and what they mean for modern data platforms. We will also give updates on our AI roadmap, including how recent acquisitions like LibreChat and Langfuse strengthen agent interfaces, observability, and evaluation, and how we are building toward AI systems that do not just answer questions, but actively help you find better ones.