Skip to content
MeetupIn personPast event

Deploying AI Agents That Perform

Event artwork for Deploying AI Agents That Perform

This event has ended

This listing is kept as an archive. Explore upcoming events to find your next one.

See upcoming events in Montreal

Supporting this event

Sponsors

BotpressMcGill AI Society
About this event

What to expect

What separates AI agent demos from production-ready systems? This evening at Botpress HQ tackles the uncomfortable truth most builders face: AI agents look impressive until they hallucinate a refund policy that costs you $10K, expose customer data, or burn through tokens in an endless loop. GenAI Montréal brings together industry leaders and the McGill AI Society to explore the architectural patterns, safety layers, and infrastructure decisions that actually make autonomous systems reliable at scale.

Why Most AI Agents Fail in Production

The session opens with Michael Masson, CTO of Botpress, sharing hard-won lessons from building the infrastructure layer for autonomous AI agents. In "We Built the Engine That Makes AI Agents Actually Work," Michael reveals where agents systematically break in real-world usage and the architectural patterns that prevent repeated failure. Rather than chasing smarter models, this talk explores why infrastructure and orchestration matter more than most teams realize—and what it really takes to move from impressive demos to systems that operate reliably.

Morgan Willis, Principal Cloud Technologist at AWS, follows with "I Don't Trust AI Agents (And Neither Should You): Building Production-Ready Architectures." Morgan walks through a layered approach to agent safety using Amazon Bedrock AgentCore and the Strands Agents SDK, covering guardrails at multiple levels—input validation, output filtering, and action approval. You'll see how multi-agent architectures can be designed so agents steer and check one another, and how observability patterns make agent behavior transparent and auditable. The reference architecture presented shows exactly how each layer catches failures before they reach production.

Production Patterns and Real-World Tooling

This isn't a high-level overview—expect concrete discussions of orchestration frameworks, security policies, prompt design patterns, cost optimization, and the monitoring infrastructure needed to make AI agents trustworthy. Whether you're working with ChatGPT, Claude, or other LLM providers, the architectural principles apply across platforms. The evening bridges academic research with industry practice, making it particularly valuable for McGill students interested in real-world AI engineering alongside developers already shipping agent systems.

Botpress works closely with McGill talent, including many interns fro

Running an AI event in Montreal?

List it here, free.

Meetups, workshops and conferences go in front of thousands of builders after a quick review.

List your event