Day 1 Online Workshops

AI agents are becoming increasingly capable of reasoning, planning, and interacting with external tools, but moving from a successful prototype to a reliable production system remains a significant engineering challenge. This session explores practical techniques for engineering and evaluating AI agents that perform consistently in real-world environments.

Discussion includes how to design agentic workflows, define meaningful evaluation metrics, and measure performance beyond simple accuracy. Topics include task completion, reasoning quality, tool selection, LLM-as-a-Judge, automated evaluation pipelines, human evaluation, regression testing, and production monitoring.

In many projects, specifications and code drift apart over time. Requirements change, documentation becomes outdated, and developers rely mainly on the code. This increases risk and makes changes harder, especially in long-lived business applications.

This talk presents a Spec-Driven Development approach in which system use cases are the central artifact. A system use case describes observable system behavior and acts as a stable contract for the application. Code is derived from these use cases instead of treating the code itself as the source of truth.

An AI agent is used as a supporting tool to generate and update code and tests from system use cases. The agent operates within a harness of skills, MCP servers, and guidelines. Through concrete examples, the session demonstrates how backend logic, database access, and UI behavior can evolve together while staying consistent with the use cases.

Finally, the session shares concrete workflows and lessons learned from real-world enterprise projects, including limitations and trade-offs of using AI in this way.

Flaky, fragile, and forgotten tests can cripple delivery pipelines and frustrate teams. This session explores why test suites often degrade into poor states and how to prevent it. Through real code examples, we’ll highlight common anti‑patterns—like duplication, excessive abstraction, and suite dependencies—and show improved approaches that make tests more focused, maintainable, and reliable. Attendees will leave with practical strategies to build automation that accelerates delivery instead of blocking it.

Building a startup MVP is fundamentally different from scaling a successful platform, maintaining a mature enterprise application, or modernizing a decades-old legacy system. Yet many organizations try to use the same software development methodology throughout their product’s entire lifecycle.

This talk explores how software development methodologies—including Lean Startup, Scrum, Extreme Programming (XP), Kanban, Disciplined Agile, DevOps, and even Waterfall—address different challenges at different stages of a product’s evolution. We’ll examine how factors such as business uncertainty, organizational scale, engineering maturity, and operational risk influence the most appropriate way of working.

Rather than asking, “Which methodology is best?”, participants will learn a practical framework for selecting and evolving development practices that fit their product, team, and organization.