Very important: This meetup is not about AI in general. It is not about AI in products. It is about using AI in the SDLC.

A good rule of thumb is: does your demo result in a code change, or is in the pipeline of an actual code change?

AI Tinkerers: Dev Tools demos focuses specifically on developer workflows with AI:

• Editors/IDEs (Cursor, Windsurf, Claude Code, Amp etc.)
• Rules/policies/prompts for code
• Issue/PR automation, branch mgmt, code review
• Agentic coding loops and MCP tools for dev tasks
• Test/gen/verify loops for code changes
• Repo context, code graphs, embeddings as inputs to coding, not product features

Suggested demo format:

1. Type (non-exhaustive, non-restrictive, but it should probably fall into one or more):

• Editor/Agent Workflow (e.g., Cursor Agent mode maximalist setup)
• Context (repo rules, style guides, MCP tools)
• Ship Loop (PRD -> issues -> tasks -> parallel execution -> PRs)
• Verification (how you gate LLM code changes: tests/static analysis/evals)
• Automation (cron/agent that opens PRs, emails status, triages tickets)

2. Code on screen (1 intro slide max).
3. Show the loop: e.g. input (PRD/issue/spec) -> context wiring -> LLM/agent steps -> verification -> git artifact (commit/PR).
4. One measured improvement (pick one): lead time, review time, test coverage delta, defects caught, tokens/cost per task.
5. One failure/anti‑pattern you hit and how you fixed it.
6. 3 ‘Try at home’ steps (flags, configs, rule snippets, CLI incantations, MCP manifest).

You should target 5 minutes of content, with a minute reserved for going over. At 6 minutes an alarm will go off. There will be 3 minutes of Q/A, and you will be able to setup at a demo table during networking/food to dive deeper.