AI talks for conferences and companies.

Spec-driven development with AI

Hand-written code and unguided prompting both end the same way: inconsistency, lost context, painful maintenance. Spec-driven development puts detailed, structured specifications at the centre and turns AI into a reliable implementation partner instead of a random code generator.

We cover why the spec is the source of truth, how to structure it for working with AI, and two tools in live demos: Google Antigravity, and Gemini CLI with the Conductor extension. You leave with concrete steps for writing good specs and using these tools day to day.

An AI agent on your own hardware: OpenClaw + DGX Spark

Local AI agents are past the experiment stage. Nothing leaves your machine, there are no token limits or monthly fees, and it all works offline.

In a live demo I show an agent in real use: its parts (OpenClaw, Ollama, Gemma 4, Google Workspace, Telegram), how it's configured, and what it does.

Spec-driven development with AI

How to write specs an AI can follow, and how to use AI as an implementation partner rather than a random code generator.

An AI agent on your own hardware: OpenClaw + DGX Spark

Local AI agents that answer to nobody but you. No token limits, no data leaving the building, works offline.

AI for small businesses

Which AI tools actually save time for small companies and freelancers. Real examples, not marketing promises.

Your topic

RAG architecture, AI in education, ethics and safety, comparing models. Tell me what you need and we'll shape it.