Code with Claude Events
anthropic's comprehensive developer workshop program teaching practical AI agent development through hands-on technical sessions. Events feature multi-track programming with specialized workshops, office hours, and competitive coding challenges focused on real-world production patterns.
Workshop Architecture
Track Structure
Founder Track: Business-focused sessions on AI integration strategies, market opportunities, and organizational transformation.
Builder Track: Technical implementation sessions covering architecture patterns, integration challenges, and production deployment.
Workshop Track: Hands-on coding sessions with specific technical deliverables and environment requirements.
Key Workshop Categories
Managed Agents Development
- "Ship your first Managed Agent": Production deployment patterns using managed-agents-api
- Session management, tool integration, and monitoring
- Python-based implementations with Streamlit interfaces
Evaluation Frameworks
- "Evals for Taste": Hill-climbing optimization for slide generation agents
- LLM-as-judge evaluation patterns
- Programmatic graders vs. human evaluation
- Docker-based evaluation environments with LibreOffice integration
Memory and Persistence
- "Agents that Remember": Long-term memory systems
- dreaming-service integration patterns
- Memory stores and contextual recall mechanisms
Competitive Development
- "Agent Battle": 45-minute optimization competitions
- Real-time performance evaluation (30-second eval loops)
- Ranking systems and peer comparison
Technical Prerequisites
Environment Setup Requirements
Core Dependencies:
- Docker (daemon running, image building capabilities)
- Node.js (v22+) with npm
- Python (3.14+) with virtual environments
- ant CLI for agent management and authentication
Security Practices:
- Proper API key management (
.envfiles, not committed) - Authentication verification before workshops
- Credit claiming and organization ID submission
Workshop-Specific Setup:
- Docker image builds for LibreOffice integration (695MB+ images)
- Python virtual environments with workshop-specific dependencies
- Repository cloning and dependency installation
- Smoke testing procedures for morning verification
Strategic Participation Patterns
Optimal Workshop Selection:
- Prioritize hands-on workshops over theoretical sessions
- Focus on unique capabilities not available elsewhere
- Choose workshops with actionable deployment patterns
- Select sessions that build on each other conceptually
Recommended Flow: "How We Claude Code" → "Ship your first Managed Agent" → "Agents that Remember" → "Evals for taste" → "Agent Battle"
Office Hours Strategy
Topic Selection Framework
Claude Platform Focus: Production deployment, pricing models, observability patterns, debugging managed agents in sandbox environments.
Research Focus: Constitutional AI methods, evaluation frameworks, alignment techniques, memory system architectures.
Claude Code Focus: Development workflows, skill integration, hook patterns, user experience optimization.
Effective Question Patterns
Actionable Production Questions:
- Observability and debugging patterns for managed agents
- Retry/idempotence patterns for agent sessions
- Cost modeling and business case development
- Migration paths from chat completions to managed agents
Technical Architecture Questions:
- Memory roadmap beyond Dreaming Service preview
- Decision trees for skills vs. MCP servers vs. sub-agents
- Production monitoring and error handling patterns
Avoid: General AGI discussions, future predictions, theoretical AI safety questions without specific implementation context.
Event Logistics
Duration: Full-day events (8am-6pm) with doors opening at 8:00am Format: Multiple parallel tracks requiring strategic session selection Materials: Repository access, API credits, organization ID requirements Follow-up: Office hours booking, express check-in systems, photo ID requirements