Phil Nash

Retrieval augmented generation (RAG) provides large language models with up to date information and helps them hallucinate less. But how does it all work beneath the covers?
In this live coding session we'll build the components of a RAG system from scratch in JavaScript. (Aside from the LLM, there probably isn't time for that!) By building our own, we'll understand vectorisation, similarity search, and the role of embedding models and vector databases. We'll then plug it all together to see our augmented bot in action.
You'll get a good grounding in the components of successful chatbots and why they work.

One of the most common issues in JavaScript code bases is that our code is too complex. As projects and their requirements evolve, complexity creeps in. Excess complexity slows progress, frustrating us as we try to keep large chunks of a program in our heads just to understand what is happening.
To conquer complexity, we must refactor! In this talk, we'll investigate how to identify unnecessary complexity in our code base using cognitive complexity as a measure, how to approach refactoring complex code, and what tools are available to help us refactor. We'll put it all into practice through a live example. By the end of the talk, you'll be excited to tackle that 1000-line monstrosity of a function in your own code base (you know the one).

Can we get rid of passwords yet? They make for a poor user experience and users are notoriously bad with them. The advent of WebAuthn has brought a passwordless world closer, but where do we really stand?
In this talk we'll explore the current user experience of WebAuthn and the requirements a user has to fulfill for them to authenticate without a password. We'll also explore the fallbacks and safeguards we can use to make the password experience better and more secure. By the end of the session you'll have a vision for how authentication could look in the future and a blueprint for how to build the best auth experience today.