Max Tkacz
The AI Agent Path to Prod: Building for Reliability
#1about 4 minutes
Why AI agents fail in production environments
AI agents often fail in production because the probabilistic nature of LLMs conflicts with the need for reliability at scale.
#2about 5 minutes
Scoping an AI agent for a specific business problem
Start by identifying a low-risk, high-impact task, like automating free trial extensions, to establish a viable solution scope.
#3about 3 minutes
Walking through the naive V1 customer support agent
The initial agent uses an LLM with tools to fetch user data and extend trials, but its reliability is unknown without testing.
#4about 4 minutes
Using evaluations to test the happy path case
Evaluations are introduced as a testing framework to run the agent against specific test cases, revealing inconsistencies even in the happy path.
#5about 4 minutes
Improving agent consistency with prompt engineering
By adding explicit rules and few-shot examples to the system prompt, the agent's tool usage and response quality become more consistent.
#6about 5 minutes
Testing for prompt injection and other edge cases
A new evaluation case for prompt injection reveals a security flaw, which is fixed by adding specific security rules to the system prompt.
#7about 6 minutes
Applying production guardrails beyond evaluations
Beyond evals, production readiness requires adding human-in-the-loop processes, custom error handling, rate limiting, and model redundancy.
Related jobs
Jobs that call for the skills explored in this talk.
Matching moments
09:56 MIN
The challenge of moving AI from demo to production
What’s New with Google Gemini?
00:05 MIN
Moving agentic AI from proof of concept to production
Building Blocks for Agentic Solutions in your Enterprise
24:42 MIN
Overcoming the challenges of productionizing AI models
Navigating the AI Revolution in Software Development
22:59 MIN
Key takeaways for building reliable LLM agents
The Limits of Prompting: ArchitectingTrustworthy Coding Agents
00:25 MIN
The challenge of building production-ready AI applications
One AI API to Power Them All
20:03 MIN
The promise and pitfalls of implementing agentic AI
Rethinking Customer Experience in the Age of AI
17:30 MIN
Live demonstration of deploying Azure resources from a prompt
Infrastructure as Prompts: Creating Azure Infrastructure with AI Agents
38:31 MIN
Previewing the "AI or knockout" conference talk
From Learning to Leading: Why HR Needs a ChatGPT License
Featured Partners
Related Videos
39:11Agents for the Sake of Happiness
Thomas Dohmke
24:22Beyond Chatbots: How to build Agentic AI systems
Philipp Schmid
29:03On a Secret Mission: Developing AI Agents
Jörg Neumann
25:32Beyond Prompting: Building Scalable AI with Multi-Agent Systems and MCP
Viktoria Semaan
24:05The Limits of Prompting: ArchitectingTrustworthy Coding Agents
Nimrod Kor
26:30Three years of putting LLMs into Software - Lessons learned
Simon A.T. Jiménez
24:11You are not my model anymore - understanding LLM model behavior
Andreas Erben
26:47The State of GenAI & Machine Learning in 2025
Alejandro Saucedo
Related Articles
View all articles
From learning to earning
Jobs that call for the skills explored in this talk.
AI Engineer Workflows & Agents (e.g. with Langdock, n8n & make)
WaveSix Labs GmbH
Intermediate
GIT
JSON
GraphQL
Microsoft Office
Evaluation and Hardening of Embedded AI Modules
Association Bernard Gregory
AI Agent Engineer (Machine Learning Engineer)
Zendesk
Berlin, Germany
Remote
FastAPI
Machine Learning
Natural Language Processing