I presented a lightning talk at “Road to SRE NEXT@Fukuoka (Hybrid Event).”

Event Overview

• Event Name: Road to SRE NEXT@福岡(ハイブリッド開催)
• Date & Time: May 9, 2025 (Fri) 19:00 - 22:00
• Venue: GMO Pepabo Inc. Fukuoka Office 7F - 2-7-21 Tenjin, Chuo-ku, Fukuoka City, Fukuoka Prefecture

Presentation: “Getting Started with LLM Tracing for AI Apps Using Langfuse”

Overview

In this lightning talk, I introduced practical examples of tracing using the Open Source LLM Engineering Platform “Langfuse.” Langfuse tracks LLM calls and related logic in detail, visualizing the processing flow through nested traces. This enables developers to accurately understand what’s happening inside their applications and quickly identify root causes of issues. Through examples of Langfuse usage in the AI App Builder “Giselle,” I shared specific benefits of implementing LLM tracing and future challenges, sharing insights from AI app development.

• Langfuse https://langfuse.com/
• Giselle https://giselles.ai/

Slides

Reflections

• The SRE NEXT organizing team’s preparation before and during the event was excellent
• GMO Pepabo Inc.’s Fukuoka office was very comfortable
• I gained new insights from other presentations, satisfying my intellectual curiosity
• The after-party allowed for off-the-record discussions that can’t be shared online, reminding me of the value of in-person participation
• Since Fukuoka is easily accessible from Kumamoto, I’d like to participate again if there are opportunities for lightning talks

That’s all from the Gemba, reporting on my lightning talk at Road to SRE NEXT@Fukuoka.