Get Production Ready: Securing and Observing LLM Applications

Description

Your LLM app works great in development, but are you ready to move it to production. What happens when users start probing it with creative prompts? When your RAG system accidentally surfaces confidential documents? When latency spikes and you have no idea why?
This workshop tackles the two most neglected aspects of production LLM systems: security and observability. You’ll learn to defend against prompt injection, implement guardrails, detect anomalies, and build the instrumentation needed to debug issues at 3am.
Through hands-on exercises with Python, you’ll build security layers and observability pipelines for a realistic RAG application — leaving with patterns you can apply to your own projects immediately.

Workshop parts

Part 1: Threat Modelling for LLM Apps
Part 2: Building Defensive Layers
Part 3: Observability Foundations
Part 4: Monitoring and Alerting
Part 5: Incident Response & Debugging

Requirements:

Comfortable in Python.

Target Audience:

Basic understanding of LLMs and RAG concepts. Software engineers building or maintaining LLM-powered applications who want to sleep better at night.