LLM-assisted development is moving quickly, and “vibe coding” conversations are happening in many organisations. In this session we’ll discuss the difference between “vibe coding” and “AI Assisted Engineering”. We will look at how the latter can be a real productivity lever teams when it’s paired with the right patterns, guardrails, and reusable building blocks. We will show how Coeo is approaching this with Alloy: our practical accelerator for building repeatable assets, speeding up delivery without compromising consistency, quality, or maintainability.
The discussion will centre around three key areas:
1) What Alloy is (and what it isn’t)
We’ll outline the problem Alloy is designed to solve, turning “how our best people do it” into reusable patterns others can leverage. We’ll discuss how standardising delivery assets (for example, configuration-driven approaches and shared notebook/utility patterns) reduces variability across projects and where AI assistance helps versus where it creates noise.
2) Vibe Coding vs AI Assisted Engineering
What is the difference between Vibe Coding and AI Assisted Engineering? Why we prefer AI Assisted Engineering, and what some of the dangerous patterns are that we see emerging with AI at the moment. We will look at how we get the most out of modern coding assistants and what you need around them (prompt patterns, validation steps, and review habits) to keep outputs production-ready.
3) Putting it into practice: Using Alloy to build out ingestion piplines in Fabric
We’ll walk through a worked-example approach: start with a blank Fabric workspace, build out our lakehouses, and ingest data from source all the way to Silver.
