Skip to content
← Back to homepage

AI Native Engineering · Developer Course

From prompt writer to AI system builder.

By Dominic Hückmann · last reviewed May 2026 · 45–60 minutes

A self-paced learning path for software developers who want to operate AI features, not just demo them — covering context budgets, Task Contracts, decomposition, evals, rollout, observability, and fallbacks.

Who is this for?

For software developers, tech leads, and AI-feature teams. This is not generic business prompting advice; it is for teams responsible for putting LLMs, coding agents, or AI-assisted product features into software safely.

Key takeaways

  • Prompting is not dead; prompt-only thinking is too small.
  • Important AI work starts with context boundaries, not clever phrasing.
  • Every risky AI output needs evidence, an eval, and a stop/fallback rule.
  • Coding agents need a repo operating system: AGENTS.md, skills, tool gates, and reviews.
  • Production AI is product work: cost, latency, UX trust, rollout, and incident response count too.

Interactive slide mode

Becoming LLM-Native

The interactive Reveal.js deck runs here as a static asset inside a sandboxed iframe. This page explains the learning curve; slide mode keeps the full details and reference tables.

Use arrow keys · Mobile works best landscape

Open deck directly

Related reading

Prompting 2026: Context Engineering →Prompt Decomposition: Break AI tasks down correctly →The LLM-native developer needs more than prompts →Open Agent Buildprints →

Sources covered

Lost in the MiddleReActTree of ThoughtsDSPyOWASP LLM Top 10NIST AI RMFAnthropic / OpenAI / Google prompting docs

Want to turn this into a team operating model?

The course points toward a practical operating model: Task Contracts, eval sets, tool gates, rollout rules, incident playbooks, and Agent Buildprints.

Ask about a project or workshop