Build Better LLM Apps

The AI Engineering Loop with Langfuse, end to end

A free 2-hour online workshop for engineers building production LLM applications.

Most teams shipping AI features today are iterating on instinct, tweaking prompts, swapping models, adjusting parameters, with no reliable way to know if the change actually made things better. That slow, blind cycle is the bottleneck between a prototype and a production-grade AI product.

This workshop gives you the tooling and workflow to close that loop. Walk away with working traces, managed prompts, and a repeatable eval workflow.

What you will build

Turn a blind app with no visibility into a functioning AI engineering loop. You'll work through a real TypeScript LLM agent — the Dad IT Support Agent — and instrument it from scratch using Langfuse.

By the end of the session you'll have:

Full trace visibility across every agent step

Prompts versioned and managed outside your codebase

A monitoring setup that catches out-of-scope requests automatically

A dataset + eval run you can reuse on any future change

Who is it for?

Engineers shipping LLM features and flying blind on what's actually happening in production

Tech leads evaluating observability tooling for their AI stack

Anyone who's asked: "How do I know if my prompt change made things better or worse?"

You should be comfortable reading TypeScript. No prior Langfuse experience needed.

What to prepare If you are interested in hands-on experience, the come prepared before the session:

Node.js 18+ installed

A Langfuse Cloud account (free) — sign up at cloud.langfuse.com

An OpenAI API key with available credits

The instructor will run through the labs with you - LIVE!

Agenda - 120 minutes

15 min: Introduction to Langfuse and the AI engineering loop

5 min: Langfuse Cloud walkthrough

10 min: Workshop environment setup

60 min: Hands-on labs — tracing, prompt management, monitoring, evals

15 min: AI engineering loop demo

15 min: Q&A and resources