SAHASRA TECHNOLOGIES
Beginner to intermediate curriculum · Claude & Codex AI-Assisted DevelopmentClaude & Codex AI-Assisted Development
A disciplined AI workflow for learning, planning, small changes, debugging, refactoring, review, tests and documentation. If you cannot explain it, do not commit it.
- Duration
- 10–14 focused classes + ongoing integration
- Per class
- 75–90 minutes
Curriculum PDF: save the complete Claude & Codex AI-Assisted Development syllabus for offline reference.
Learning objective
Use Claude and Codex to support learning and small verified changes while retaining technical ownership.
Recommended class plan
10–14 focused classes + ongoing integration · 75–90 minutes per class
1 focused class + responsible AI practice throughout
Technologies
- Claude
- OpenAI Codex
- AI-assisted development workflow
Tools
- Claude
- Codex
- VS Code workflows
- Git
- Compiler
- Test runner
Curriculum modules
Basic Foundations → Practical Skills → Guided Application
- 1
AI as a Learning Partner
Explanations, examples, comparisons, exercises and knowledge checks.
- 2
Context & Prompting
Provide requirements, constraints, project context and ask for an approach before code.
- 3
Small Verified Changes
Generate a limited change, read every file, review the diff, compile and test.
- 4
Debugging, Refactoring & Review
Investigate errors, improve code, draft tests and review security and correctness.
- 5
Responsible Vibe Coding
Ask → generate → understand → review → test → improve → commit.
- 6
AI + Git Workflow
Requirement → branch → approach → small change → diff → compile → test → review → commit → PR.
Beginner topics
Start Here and Build Comfort
- Ask for beginner explanations
- Generate exercises without answers
- Explain unfamiliar code
- Provide focused context
Intermediate topics
Progress After the Basics Are Clear
- Review diffs
- Debug step by step
- Generate and verify tests
- Refactor small scopes
- AI-assisted code review
- Git-integrated workflow
Practical learning
Practise, Submit, Build and Explain
Practical exercises
- Improve a vague prompt
- Review a generated method
- Find an AI-introduced bug
- Explain a diff before committing
Assignments
- Complete one AI-assisted feature with evidence
- Create a personal verification checklist
Mini projects
- AI-Assisted CRUD Feature Under Review
Industry habits throughout the track
Terminology, Debugging and Responsible AI
Industry terminology
- Context
- Constraint
- Acceptance criteria
- Diff
- Hallucination
- Refactor
- Regression
- Ownership
Debugging exercises
- Challenge an incorrect AI diagnosis
- Use logs and tests to verify a suggestion
- Reject an unnecessary rewrite
AI learning exercises
- Learning prompts
- Coding-practice prompts without solutions
- Line-by-line explanations
- Interview simulation
- Test and documentation review
Topics intentionally excluded
Advanced Topics Come Later
Quality fundamentals are more valuable than an overloaded syllabus. These topics are outside this track:
- Blind application generation
- Committing unexplained code
- Sharing credentials
- Autonomous production changes
- Replacing fundamentals with prompts
Recommended next track