Training & Skill Development

AI-assisted development is a skill, not an instinct. Giving developers access to AI tools without training is like giving a team a new programming language and expecting production code by Monday. This section defines the training curricula, skill assessment frameworks, learning paths, and certification programs that build AI-assisted development capabilities systematically across all engineering roles.

Training Philosophy

The AEEF training approach is built on four principles:

1. Progressive Depth: Training moves from literacy to proficiency to mastery, with each level building on the previous one
2. Role-Specific: Different roles require different skills; one-size-fits-all training wastes time and misses critical needs
3. Practice-Based: AI-assisted development is a practical skill. Training MUST include hands-on exercises, not just theory
4. Continuously Updated: AI tools evolve rapidly. Training materials MUST be reviewed and updated quarterly

info

Training is not a one-time event. It is a continuous program. Organizations MUST budget for ongoing training, not just initial onboarding. The AI tools and best practices that are current today will be outdated within 12 months.

Skill Assessment Framework

Before designing training, organizations MUST assess current skill levels to identify gaps and prioritize investment.

AI-Assisted Development Skill Matrix

Skill Domain	Beginner (1)	Competent (2)	Proficient (3)	Expert (4)
AI Tool Operation	Can access and use basic features	Uses all standard features effectively	Configures tools for optimal performance	Evaluates and customizes tools for team needs
Prompt Engineering	Writes basic prompts with simple instructions	Constructs structured prompts with context	Designs multi-step prompts with constraints and examples	Creates reusable prompt templates, mentors others
Context Preparation	Includes minimal context	Selects relevant code and documentation	Curates optimal context packs for different scenarios	Designs context strategies for the team
Output Evaluation	Accepts/rejects based on obvious errors	Evaluates correctness, style, and edge cases	Assesses security, performance, and architectural fit	Defines evaluation criteria for the team
Iterative Refinement	Makes one attempt, then codes manually	Refines 2-3 times with improved prompts	Systematically refines using the DCRI model	Optimizes refinement workflows for efficiency
Quality & Security	Aware that AI code needs review	Applies standard review practices to AI code	Applies enhanced review for AI-specific risks	Designs quality gates and review processes
Workflow Design	Uses AI ad-hoc, no consistent pattern	Follows team workflow patterns	Designs and improves workflow patterns	Architects organizational workflow standards
Mentoring & Teaching	N/A	Can explain basic AI tool usage	Mentors team members in AI-assisted practices	Trains across the organization, contributes to curriculum

Assessment Process

1. Self-Assessment: Developers complete the skill matrix self-assessment (15 minutes)
2. Peer Validation: AI Champions validate self-assessments through brief observation (optional but RECOMMENDED)
3. Gap Analysis: Compare individual assessments to role-level expectations
4. Training Prescription: Assign learning path modules based on identified gaps

Assessment Cadence:

• Initial assessment: During onboarding or program launch
• Reassessment: Every 6 months
• Targeted assessment: After completing a training module or certification

Training Curricula

Tier 1: AI Literacy Fundamentals (All Engineering Staff)

Duration: 4 hours (can be split across 2 sessions) Format: Instructor-led with hands-on exercises Prerequisites: None

Curriculum:

Module	Duration	Content
1.1 AI-Assisted Development Landscape	45 min	What AI tools do, how they work (conceptually), industry adoption trends, organizational strategy
1.2 Quality and Risk Awareness	45 min	Documented quality risks (1.7x issues, 2.74x vulnerabilities), why review matters, the AEEF quality framework
1.3 Hands-On Tool Introduction	90 min	Basic tool usage, simple code generation, accepting/rejecting suggestions, basic prompting
1.4 Organizational Standards	60 min	Approved tools, prompt repository, workflow standards, quality gates, where to get help

Tier 2: Practitioner Skills (All Developers)

Duration: 12 hours (delivered over 2-3 weeks) Format: Blended (instructor-led workshops + self-paced exercises + pair programming) Prerequisites: Tier 1 completion

Curriculum:

Module	Duration	Content
2.1 Prompt Engineering	2 hours	Structured prompting, context inclusion, constraint specification, prompt patterns
2.2 Context Preparation	2 hours	Context types, preparation checklist, context packs, context window management
2.3 The DCRI Workflow	2 hours	Task decomposition, context preparation, iterative refinement, integration practices
2.4 Output Evaluation	2 hours	Code review for AI output, security checks, quality assessment, common AI failure modes
2.5 Test Generation	2 hours	AI-assisted test creation, assertion validation, coverage analysis, test strategy design
2.6 Practical Workshop	2 hours	End-to-end feature development using AI-assisted workflow, peer review

Tier 3: Advanced Practitioner (Senior Engineers, AI Champions)

Duration: 8 hours (delivered over 2 weeks) Format: Workshop-based with real project exercises Prerequisites: Tier 2 completion + 3 months of AI-assisted development experience

Curriculum:

Module	Duration	Content
3.1 Advanced Prompt Design	2 hours	Multi-step prompts, chain-of-thought, few-shot patterns, domain-specific prompting
3.2 Workflow Architecture	2 hours	Designing team workflows, automation library development, prompt repository curation
3.3 Quality Gate Design	2 hours	Designing AI-specific quality gates, static analysis configuration, security scanning
3.4 Mentoring & Knowledge Transfer	2 hours	Teaching AI-assisted development, peer coaching, community of practice facilitation

Tier 4: Leadership & Strategy (Engineering Managers, CoE Staff)

Duration: 6 hours Format: Seminar with case studies Prerequisites: Tier 1 completion + management role

Curriculum:

Module	Duration	Content
4.1 Managing AI-Augmented Teams	2 hours	Estimation changes, capacity planning, performance evaluation
4.2 Change Management	2 hours	Change management strategies, resistance management, communication
4.3 Metrics & ROI	2 hours	Metrics framework, ROI calculation, stakeholder reporting

Learning Paths

Path A: Developer (IC)

Tier 1 → Tier 2 → [3 months practice] → Tier 3 (optional) → Ongoing quarterly updates

Path B: Engineering Manager

Tier 1 → Tier 4 → Tier 2 (abbreviated, hands-on awareness) → Ongoing quarterly updates

Path C: AI Champion

Tier 1 → Tier 2 → [3 months practice] → Tier 3 → Champion-specific coaching → Ongoing monthly CoP

Path D: QA Engineer

Tier 1 → Tier 2 (with QA-specific exercises) → [3 months practice] → Tier 3 Module 3.3 → Ongoing updates

Path E: New Hire Onboarding

Tier 1 (Week 1) → Tier 2 (Weeks 2-4) → Buddy pairing with AI Champion (Months 1-3)

tip

New hires who have AI-assisted development experience from previous roles SHOULD take a placement assessment and may skip Tier 1 if they score at the Competent level or above on the skill matrix. They MUST still complete the Organizational Standards module (1.4) regardless of prior experience.

Certification Program

Certification Levels

Level	Name	Requirements	Validity
Foundation	AEEF AI-Assisted Developer	Tier 1 + Tier 2 completion + skills assessment at Competent (2) or above	1 year
Practitioner	AEEF AI-Assisted Development Practitioner	Foundation + 6 months experience + skills assessment at Proficient (3) or above	1 year
Expert	AEEF AI-Assisted Development Expert	Practitioner + Tier 3 completion + demonstrated contribution to prompt repository and workflows	2 years
Champion	AEEF AI Champion	Expert + Tier 3 Module 3.4 + active Champion role for 6+ months	2 years

Certification Maintenance

• Foundation and Practitioner certifications require annual renewal through a brief reassessment and completion of quarterly update modules
• Expert and Champion certifications require biennial renewal plus demonstrated ongoing contribution
• Certifications automatically expire if the holder does not complete required renewal activities
• Expired certifications can be reinstated through reassessment without repeating full training

warning

Certification MUST NOT be used as a gatekeeping mechanism that prevents developers from using AI tools. All developers with Tier 1 training may use approved AI tools. Certification is a recognition of skill level, not a prerequisite for tool access.

Training Delivery Standards

Instructor Requirements

• Tier 1 and Tier 2 training MAY be delivered by AI Champions who have completed Tier 3
• Tier 3 training MUST be delivered by CoE staff or certified external trainers
• Tier 4 training MUST be delivered by CoE leadership or senior engineering leadership
• All instructors MUST have active, hands-on experience with AI-assisted development (not just theoretical knowledge)

Training Environment

• All hands-on training MUST use the organization's approved AI tools in a sandboxed training environment
• Training environments MUST NOT have access to production code or data
• Training exercises SHOULD use realistic but fictional codebases that represent the organization's technology stack
• Training environments MUST be available for self-study after formal sessions

Training Effectiveness Measurement

Training effectiveness MUST be measured and reported to the CoE:

Metric	Target	Method
Training completion rate	95% within first 6 months	LMS tracking
Knowledge assessment pass rate	85% first attempt	Post-training assessment
Skill level improvement	Average +1 level within 3 months	Skill matrix reassessment
Developer satisfaction with training	4.0+ out of 5.0	Post-training survey
On-the-job application rate	80% using learned techniques within 2 weeks	Follow-up survey + telemetry

Continuous Learning

Beyond formal training, organizations MUST support continuous learning:

• Monthly AI Tool Updates: Brief sessions (30 minutes) covering new features, updated best practices, and lessons learned
• Community of Practice: Regular forums where developers share techniques, prompts, and experiences
• External Learning Budget: Allocate budget for developers to attend AI-related conferences and courses
• Internal Knowledge Base: Maintain a searchable repository of AI-assisted development tips, tricks, and case studies, linked to the prompt repository
• Pair Programming with AI: Encourage regular pair programming sessions where developers collaborate on AI-assisted tasks, learning from each other's approaches