Full Claude Mastery
From first prompt to working agents, the whole Claude toolkit in one live workshop.
Who it’s for: Professionals, founders, lawyers and students. No coding needed.
- FoundationsChoosing the right model and plan for the job · prompt and context engineering (roles, structure, examples, long documents) · getting answers you can actually check
- Everyday work productsProjects and knowledge bases · Artifacts for building small apps, dashboards and documents inside a chat · analysing PDFs, spreadsheets and contracts · writing, briefs and decks
- AgentsHow tool use works · connecting Claude to your email, calendar, docs and databases with MCP · building a simple multi-step agent · Skills: reusable instructions that make Claude work your way
- Claude CodeThe terminal agent: reading a project, making changes, running tests, and why it’s useful even if you don’t code (scripts, data clean-up, small websites)
- AutomationRecurring and scheduled tasks · triggering Claude from n8n, Zapier and webhooks · build a daily briefing, an inbox-triage flow and a report-writer
- Safe and reliablePrivacy settings and what never to paste · spotting hallucinations · evaluating outputs · keeping costs under control
You leave with: Your own personal agent plus three automations running.
Live online workshop · replay access included. Batch dates are announced to enrolled students.
Independent workshop, not affiliated with or endorsed by Anthropic. “Claude” is a trademark of its owner.
OpenClaw Workshop
Run your own always-on personal AI agent, and keep it locked down.
Who it’s for: Builders and curious professionals comfortable following setup steps. No prior agent experience needed.
- What it isOpenClaw is an open-source, self-hosted agent that you talk to through the chat apps you already use. When it’s the right tool, and when it isn’t
- Safe setupRunning it on an isolated machine, VM or container · least-privilege accounts · handling API keys and secrets properly
- Models and channelsConnecting and switching model providers · controlling cost · pairing Telegram and WhatsApp-style channels and allow-listing who may talk to it
- Skills and memoryAdding skills · long-term memory · scheduling background tasks
- Build three agentsA morning briefing · an inbox-and-calendar assistant · a research-and-report agent
- HardeningPrompt-injection awareness · permissions and approvals · logging · a kill-switch · what you should never give an agent access to
You leave with: A locked-down personal agent running on your own device or server.
Live online workshop · replay access included. Batch dates are announced to enrolled students.
Independent workshop, not affiliated with the OpenClaw project. An agent with computer access carries real risk; we teach a safety-first setup, but you remain responsible for your own systems.
Grok Bots Workshop
Build bots on xAI’s Grok models, from a chat assistant to a tool-using agent.
Who it’s for: Anyone who wants a working bot online by the end of the session. No coding needed.
- Grok and its APIThe Grok model family · API keys, limits and cost awareness
- Design the botDefining the job, persona and guardrails before writing a single prompt
- ToolsFunction/tool calling so the bot can look things up and take actions · using live-search and real-time features where the API offers them
- DeployPutting it on Telegram, a web chat on your site, and a simple API endpoint
- OperateModeration, abuse limits and logging · testing with a small evaluation set before you go live
You leave with: A deployed Grok-powered bot with at least two tools.
Live online workshop · replay access included. Batch dates are announced to enrolled students.
Independent workshop, not affiliated with or endorsed by xAI. Features depend on xAI’s API at the time of the session.
Train Your Bots
Turn a generic bot into an expert in your domain, without training a model from scratch.
Who it’s for: Anyone building a bot, assistant or agent for a specific field, team or business.
- What “training” really meansInstructions, examples, retrieval, memory and fine-tuning, and which one to reach for first (usually not fine-tuning)
- Your knowledge baseCleaning documents, chunking and metadata so a bot can find the right passage
- Retrieval with citationsGrounding answers in your documents (RAG) and showing sources so people can verify
- Behaviour that sticksSystem prompts and few-shot examples · memory and personalisation
- Measure itWriting a test set, scoring answers, tracking regressions every time you change something
- Improve itFeedback loops that turn corrections into better behaviour · when fine-tuning is worth the cost · guardrails: scope, refusal and hand-off to a human
You leave with: A domain-expert bot with a 30-question test set and a score you can improve.
Live online workshop · replay access included. Batch dates are announced to enrolled students.
Use sample or public data in the workshop, never confidential or client material.