Saturday 25 July 2026 Sunday 26 July 2026
Pending14:00–18:00 Europe/London
£250
per person · online via Zoom
Mastering Claude Code — Your team is using AI. But are they using it well? Spend a day turning unpredictable "vibe coding" into a reliable, reviewable, spec-driven workflow with Claude Code — from configuration, through OpenSpec, to a team of specialised agent personas. Co-presented by Emmz Rendle and Dylan Beattie.
Your team is using AI — are they getting the most out of it?
The problem with vibe coding
AI coding assistants are powerful — but without structure, they produce inconsistent results. Context gets lost between sessions. Requirements drift. Code reviews become guesswork. You end up fixing AI mistakes instead of shipping features.
There's a better way
Spec-driven development creates a shared understanding between your engineers and the AI — before any code is written. The result: faster development, fewer rework cycles, and code that actually matches what you wanted.
This workshop teaches it
One workshop, two formats. The same curriculum is delivered either as a private one-day business session for your team, or as two consecutive 4-hour public sessions. Hands-on, adapted to your stack — you'll leave with working setups you can commit that day.
Curriculum
The same curriculum is delivered either as one full day (private/business) or as two consecutive 4-hour sessions (public).
Part 1 — Configure Claude Code for real work
- Write a personal system prompt that shapes Claude Code across all your projects
- Create
CLAUDE.mdproject memory so Claude has instant context every session - Build custom slash commands for the workflows you run every day
- Create reusable skills Claude invokes automatically for recurring tasks
- Connect external tools via MCP (GitHub, databases, and more)
- Set up hooks that auto-format, lint, and type-check without manual steps
Part 2 — Spec-driven development with OpenSpec
- Why spec-driven development beats vibe coding for reliable, reviewable output
- The OpenSpec workflow: proposal → specs → design → tasks → implementation
- Writing Given/When/Then specifications that guide AI implementation precisely
- Hands-on: build a real feature end-to-end with OpenSpec
Part 3 — A team of agent personas
- Build specialised agent persona files — planner, implementer, reviewer, tester, refactorer
- Delegate work to subagents and coordinate personas through files and specs
- When to reach for a multi-agent workflow — and when not to
- Hands-on: drive a feature build with a team of agent personas
- (With Claude Max: scaling to concurrent sessions and git worktrees for isolated parallel work.)
All exercises are adapted to your team's language, framework, and tooling.
What you'll walk away with
- A personalised Claude Code setup committed to your repository and ready to use
- Custom commands and skills tailored to your workflows
- The ability to write specifications that guide AI implementation precisely
- A repeatable OpenSpec workflow for shipping features with Claude Code
- A set of agent personas adapted to your projects
- Confidence to use AI agents on real production work — not just toy examples
- Knowledge of when to use AI agents vs. when not to
- 30 days of follow-up support via email
About the Co-presenters
Emmz Rendle
Emmz has been writing software professionally for 30+ years, and has spent the last year using Claude Code for real production work.
A keynote speaker with talks at NDC, GOTO, and others, Emmz brings a developer-to-developer perspective: no hype, no fluff, just workflows that actually work on real codebases.
Dylan Beattie
Dylan Beattie is a software consultant, developer, and speaker with over 25 years of experience building and shipping software. Known for talks at NDC, JFokus, and dozens of other conferences worldwide, Dylan has been using AI-assisted workflows in production systems and writes about engineering practices at dylanbeattie.net.
Dylan runs Ursatile Ltd, through which workshop payments are collected and invoiced (VAT registered).
Mastering Claude Code is co-presented by Emmz Rendle and Dylan Beattie — two practitioners using Claude Code on real production work.
Watch: “How I Tamed Claude” — NDC London 2026
Formats & Pricing
Public — two-day workshop
Two consecutive 4-hour sessions, online via Zoom. £250 per person.
Minimum 4 participants for a date to run. Register your interest in a date and we'll confirm when it reaches the minimum — no payment at the interest stage.
Business — one-day workshop
One full day (8 hours) for your engineering team, online or in-person at your premises, fully adapted to your stack. Suitable for 8–20 engineers. Pricing on request — get in touch.
Subscription requirement
Attendees need a Claude Pro subscription (or Anthropic API key). Claude Max is recommended to get the most from the agent-persona and parallel-workflow material.
Book for Your Team
Running a private workshop for your engineering organisation? We'll adapt the full curriculum to your stack, tooling, and codebase — and deliver it as a single intensive day, online or on-site.
Whether you're rolling out AI tooling across a team for the first time, or you want to level up engineers who are already experimenting with Claude Code, the business workshop gives everyone the same shared vocabulary, the same reliable workflow, and working setups they can commit before the session ends.
Get in touch to discuss dates, group size, and format. Pricing is on request.
Upcoming Public Dates
Book for Your Team
Frequently Asked Questions
Do I need to know Claude Code already?
No — we start from installation and configuration. You need basic command-line experience and a Claude Pro subscription (Max recommended).
What's the difference between the public and business workshops?
Same curriculum, different delivery. The public workshop runs online over two consecutive 4-hour sessions and is booked per person via expression of interest. The business workshop is a private one-day session for your team, online or on-site, adapted to your stack, with pricing on request.
What does "minimum 4 participants" mean?
A public date needs at least 4 registered participants to run. If a date doesn't reach 4, we offer registered participants a refund or transfer to the next date, with at least 2 weeks' notice.
Is this for beginners or experienced developers?
Experienced developers who want to use AI tools more effectively. You should be comfortable with Git, the command line, and your language. It's about workflow and tooling, not programming fundamentals.
What tech stack does the workshop use?
Public workshops use TypeScript/Node.js for examples; private business workshops are fully adapted to your stack.
How does "registering interest" work?
It's not a commitment or payment. It tells us you're interested in a date so we know when one is likely to run; when a date reaches the minimum we contact everyone with how to confirm and pay.
Who runs the workshop?
It's co-presented by Emmz Rendle and Dylan Beattie — two practitioners using Claude Code on real production work.
Can I get a VAT invoice?
Yes — workshop payments are collected and invoiced through Ursatile (Dylan Beattie's company, VAT registered) once a date is confirmed.
Is there follow-up support?
Yes — 30 days of email follow-up support is included.