Track B - Session 10
Observability for AI Troubleshooting, Performance, and Cost Optimization
Session Overview
Operating AI services brings unique challenges such as performance degradation, rising costs, and quality management. This session will demonstrate how Splunk enables full observability across the AI lifecycle—from GPU infrastructure to LLM services—helping enterprises quickly diagnose issues and maximize operational efficiency.
Key Topics:
- Monitoring AI infrastructure (GPU)
- Managing AI service (LLM) performance
- Quality evaluation and improvement
- Observability for Agentic AI
- Cost analysis and optimization
Speaker
In Pil-gyo
Observability Advisor
Splunk
In Pil-gyo is an Observability Advisor at Cisco, where he leads business development initiatives for observability solutions.