events

Virtual: Shift-Left Performance Engineering for Agentic AI with CI/CD

Kandasamy Selvaraj presents a virtual session for the Innovate QA Meetup Group on shift-left performance engineering for agentic AI. The talk argues that response time, error rate, and throughput are insufficient for agent pipelines, then p

Starts

2026-09-23T16:00:00+00:00

Organizer

Innovate QA Meetup Group

Location

Online event

Kandasamy Selvaraj presents a virtual session for the Innovate QA Meetup Group on shift-left performance engineering for agentic AI. The talk argues that response time, error rate, and throughput are insufficient for agent pipelines, then proposes trace-based SLOs covering tool and LLM call counts, chain depth, retries, delegation errors, timeouts, and average tokens per trace. It shows how Langfuse trace data can feed CI/CD promotion gates, including a demo and a starting threshold for flagging traces. The session is aimed at QA and performance engineering practitioners working with agentic systems.