Prompt Engineering
Prompt Architecture Workshop
Composable prompt blocks, evaluation sets, and reviewer workflows for teams that outgrew ad-hoc ChatGPT tabs.
- Duration
- 2 weeks · intensive
- Format
- Workshop sprint
- Level
- Foundational
- Tuition (informational)
- ₩450,000
Program narrative
We treat prompts like versioned interfaces. You will design block libraries, negative test cases, and lightweight reviewer shifts that keep tone aligned with brand risk appetite.
What is included
- · Prompt diff visualizations for reviewers
- · Negative test harness with tagged failures
- · Style guide integration for multilingual teams
- · Sampling parameter matrix with cost notes
- · Human calibration games for borderline intents
- · Template for deprecating unsafe prompt blocks
- · Office hours on negotiation with legal stakeholders
Outcomes you can demo
- · Publish a prompt library with semantic versioning
- · Ship an evaluation set others can extend without breaking IDs
- · Run a reviewer calibration session with documented agreement
Mentor of record
Zara Iqbal
Conversation designer turned prompt architect for regulated chatbots.
Participant questions
Is this only for chatbots?
Patterns apply to copilots and internal assistants; we use chat framing for clarity.
Coding depth?
Light scripting for harnesses; no full-stack app build.
Limitation?
We do not teach jailbreak offense beyond defensive awareness; offensive tooling is out of scope.
Recent participant notes
“Prompt Architecture Workshop finally gave us reviewer shifts that legal did not dread.”
“Negative test harness caught a tone drift before marketing saw it.”