This talk is a continuation of the one I gave at LRUG in March 2025.
Later that year I really started to reap the benefits of DDD and Ports and Adapters: refining the programmer tests (no more Docker for unit tests!) and bringing the ubiquitous language to life,
which made writing tests cheap, and more importantly, fun.
Towards the end of 2025, I gave in to FOMO and started exploring AI-assisted programming. I was particularly interested to see if it was now possible to prompt the LLM entirely in the ubiquitous language and produce well-structured code.
Is the future of programming DDD’s ubiquitous language?
When the talks come to an end we’ll decamp to a local pub for some food, some
drinks and some chat with your fellow attendees.
Of course, even though this is the socialising part and seems more
informal, please remember that still we consider it to be a part of the
meeting and covered by our code of conduct.
Please use the street address and not just the postcode if searching on Google Maps
On arrival you will then be registered with building security and will be permitted up to the 3rd floor via the lifts, where we will greet you and assign you a visitor’s badge, which you should wear at all times whilst in the building. You will be asked to present a photo ID so please come prepared.
The venue has a hard limit of 100 people. If you register and realise you
can’t come, please use tickettailor to give up your place so we can someone
else come in your place. We might be able to let in people on the night
who haven’t registered, but we can’t guarantee it.
In this talk I’ll compare today’s AI moment to past tech S-curves, and argue it’s not only a toolset change that’s required (learning to code agentically) but a deeper change in how we work.
I’ll cover the mindset and capability shifts this demands: outcome ownership, faster feedback loops, more autonomy. I’ll cover what engineers can do to meet the moment.
I’ll also share how this is changing the way I think about hiring and scaling teams: what signals matter now, and how we’re adapting interview loops to find people who thrive in an agentic world
A brief guide to building out a retrieval augmented generation system on Rails.
We’ll cover some of the basic (chunking, embeddings) and some not so basic things (why bother re-ranking). You’ll come away with an understanding of the different moving parts of a RAG system and some tactical recommendations of gems and services to use.
When the talks come to an end we’ll move to a nearby pub for some food, some
drinks and some chat with your fellow attendees.
Of course, even though this is the socialising part and seems more
informal, please remember that still we consider it to be a part of the
meeting and covered by our code of conduct.
The venue has a hard limit of 80 people. If you register and realise you
can’t come, please use let us know via TickeTailor so we can give your place
to someone in your place. We might be able to let in people on the night
who haven’t registered, but we can’t guarantee it.
At BBB, we use feature flags extensively in one of our Rails applications.
Removing fully rolled-out feature flags is always a chore that’s easy to ignore and eventually becomes a source of tech debt. In the talk, I’ll show how we use GitHub Actions and GitHub Copilot to automatically generate PRs that remove obsolete flags and their associated code paths.
A lightning talk about how we are replacing an all-consuming User object with domain-specific roles, delegated types and a smattering of metaprogramming – presented with a 1950s B-movie vibe
Once all that is over, we will have plenty to talk about so we’ll head on
out to a nearby pub for some food, some
drinks and some chat with your fellow attendees.
Of course, even though this is the socialising part and seems more
informal, please remember that still we consider it to be a part of the
meeting and covered by our code of conduct.
The venue has a hard limit of 100 people. If you register and realise you
can’t come, please use let us know via TickeTailor so we can give your place
to someone in your place. We might be able to let in people on the night
who haven’t registered, but we can’t guarantee it.
Rails lost its way during versions 4, 5, and 6 - becoming overly complex and slow as it leaned heavily into JavaScript. But the story doesn’t end there. With Rails 8.1 introducing features like Stimulus, Action Cable, and Solid Cable, the framework is returning to its roots: simplicity, speed, and developer joy.
Join us for a deep dive into why Rails has its mojo back and what this means for the future of Ruby development
In this talk, Sebastian will dive into RSpec as a craft — showing how thoughtful test design can make refactoring less terrifying. You’ll see real-world examples of transforming low-signal specs into readable, behaviour-driven tests, and learn how Junie, an IDE-native AI assistant, can act as a second pair of eyes to improve naming, structure, and guarantees.
This is a practical, opinionated look at using AI to support test quality—while keeping human judgment firmly in control.
Why join?
Many test suites pass—and still make refactoring terrifying. Why?
Explore clarity, intent, and refactor safety in RSpec.
See how AI can enhance craftsmanship without replacing developers.
Once all that is over, we will have plenty to talk about so we’ll head on
out to a nearby pub for some food, some
drinks and some chat with your fellow attendees.
Of course, even though this is the socialising part and seems more
informal, please remember that still we consider it to be a part of the
meeting and covered by our code of conduct.
The venue has a hard limit of 80 people. If you register and realise you
can’t come, please use let us know via TickeTailor so we can give your place
to someone in your place. We might be able to let in people on the night
who haven’t registered, but we can’t guarantee it.
