Plan du cours
Block 1 — Shared Foundations (Days 1–2)
Day 1 — Morning: The Human Factor in AI Adoption
• Trust / reliance calibration: when to use AI, when to stop.
• Team agreement structure (trigger / action / evidence / owner).
• Prompt Curator role: validation, decision, sign-off. AI incident response plan.
Day 1 — Afternoon: Constraints, Risks and Compliance
• Real LLM capabilities — prompt risk vectors: injection, data leakage, hallucinations.
• Legal framework: GDPR, EU AI Act — sector standards (DICOM, HL7, HIPAA).
• Practical exercise: translate a domain standard into a prompt guardrail.
Day 2 — Morning: Technical Architecture of Prompts
• Agent architecture: memory, context, goals — from a prompt design perspective.
• API integration and domain data sources, multi-agent and prompt chaining.
Day 2 — Afternoon: Enterprise Prompt Anatomy
• The 6 layers: Role / Context / Constraints / Domain Standards / Format / Examples.
• Prompt hierarchy: System (org-wide) — Domain (team) — Task (individual).
• Demo: deconstruct a naive prompt, rebuild it. Team brief for Days 3–5.
Block 2 — Co-Construction Workshops (Days 3–4–5)
Day 3 — Discovery and Standards Audit
- Parallel team workshops: Architects, Domain-Specific Devs, Back-End, QA.
- Mapping enterprise standards and constraints — identifying cross-team conflicts.
- Day 3 Deliverable: Standards Map + impact/effort priority matrix.
Day 4 — Convention Design and Template Construction
- Naming conventions, versioning, tag system (team, domain, target tool).
-
Building first validated templates: TypeScript DICOM, code review, QA tests, API
documentation. - Day 4 Deliverable: 4+ operational templates + conventions guide.
Day 5 — Library Assembly, Governance and Official Handover
- Library organization, GitHub Copilot / Cursor / internal LLM API integration.
- Prompt Curator role, quality metrics, team rituals, 30-day deployment plan.
- Final Day 5 Deliverable: Documented Library v1.0 + Governance Charter + 30-Day Plan.
Pré requis
- Having completed at least one AI training (introductory or advanced).
- Technical profiles: development experience in the company's stack.
- Management profiles: basic familiarity with AI tools (ChatGPT, Copilot, etc.).
- Company commitment: active participation of team leaders in Days 3–5.
- Prior provision: existing standards documentation (README, coding guides).
Target audience
- Software architects
- Developers (domain-specific, back-end, front-end)
- QA engineers / Code technicians
- Team leaders and middle managers
- IT managers, decision-makers and AI project leads
Nos clients témoignent (2)
J'ai acquis des connaissances sur la bibliothèque Streamlit en Python et je vais certainement essayer de l'utiliser pour améliorer les applications de mon équipe qui sont actuellement développées avec R Shiny.
Michal Maj - XL Catlin Services SE (AXA XL)
Formation - GitHub Copilot for Developers
Traduction automatique
Formateur capable d'ajuster le niveau du cours pendant la formation pour correspondre à notre niveau de compréhension sur le sujet, afin que nous puissions acquérir des connaissances plus utiles qui nous aideront davantage à maîtriser les outils dans notre travail quotidien.
Tatt Juen - ViTrox Technologies Sdn Bhd
Formation - Intermediate GitHub Copilot
Traduction automatique