Why Your AI Agent Will Fail In Production

Join us for a free live session ahead of the announcement of Building Reliable AI Agents, Ali Parandeh's upcoming Manning liveProject series.

This is an early, unfinished look at the material, straight from someone currently writing it.
In this talk, Ali will cover why AI agents that work perfectly in a demo often fail once they're serving real traffic: tool calls that fail silently, retrieval that breaks under load, and evaluation that gave false confidence before anyone noticed.

This session is designed for anyone building with LLMs beyond notebook experiments, whether you're shipping your first agent or trying to work out why the one you shipped keeps misbehaving.

• Date: Monday 10 August 2026
• Time: 6:30pm to 7:30pm (1 hour)
• Location: Online (Free)

We'll cover:

• Why agents that pass every test still fail with real users
• Failure patterns in tool-calling and multi-step workflows
• Evaluation that catches problems before production, not after
• Guardrails worth having versus guardrails that just slow you down
• A preview of the upcoming Building Reliable AI Agents Series (Manning, 2026)

No coding setup required, this is a talk with live examples.

Speaker
Ali Parandeh is a Chartered Engineer (CEng) with the UK Engineering Council and author of Building Generative AI Services with FastAPI (O’Reilly, 2025). With a career spanning engineering consultancy, global tech firms, and AI start-ups, he has led teams to design and deploy scalable AI-powered products across transport, finance, education, healthcare, and manufacturing.

Certified by both Microsoft (Azure Developer) and Google (Professional Data Engineer), Ali is recognised for making complex AI concepts practical and accessible. He delivers training for the Institution of Mechanical Engineers and runs Build Your AI, a consultancy focused on getting safe AI adoption right in heavy industry and engineering organisations. He speaks regularly at industry events and is known for pairing real technical depth with practical applications.

Agenda

• 6:30-6:35pm: Introductions & conceptual overview
• 6:35-7:15pm: Talk with live examples
• 7:15-7:30pm: Q&A and closing

Spaces are limited. Registration is free but required to secure your seat.
Register at: https://luma.com/ecyxsgcn

You can learn more about the upcoming series at: https://buildinggenai.com