Event 1

Summary Report: AWS Cloud Mastery Series #1 - AI/ML/GenAI on AWS

Event Information

Event Name: AWS Cloud Mastery Series #1 - AI/ML/GenAI on AWS
Date: Saturday, November 15, 2025
Time: 8:30 AM – 12:00 PM
Location: AWS Vietnam Office

Event Objectives

  • Understand the AI/ML landscape in Vietnam
  • Learn about AWS AI/ML services, particularly Amazon SageMaker
  • Explore Generative AI capabilities with Amazon Bedrock
  • Gain hands-on experience through live demonstrations
  • Network with AWS professionals and fellow participants

Event Schedule

8:30 – 9:00 AM | Welcome & Introduction

  • Participant registration and networking
  • Workshop overview and learning objectives
  • Ice-breaker activity
  • Overview of the AI/ML landscape in Vietnam

9:00 – 10:30 AM | AWS AI/ML Services Overview

Amazon SageMaker – End-to-end ML Platform

The session covered comprehensive aspects of Amazon SageMaker:

  • Data Preparation and Labeling

    • Tools for data preprocessing
    • Built-in labeling workflows
    • Integration with data sources

  • Model Training, Tuning, and Deployment

    • Distributed training capabilities
    • Hyperparameter optimization
    • One-click deployment options
    • A/B testing and model monitoring

  • Integrated MLOps Capabilities

    • Model registry and versioning
    • CI/CD for ML workflows
    • Automated model retraining
    • Performance monitoring and drift detection

Live Demo: SageMaker Studio Walkthrough

  • Hands-on demonstration of SageMaker Studio interface
  • End-to-end ML workflow example
  • Best practices for model development

10:30 – 10:45 AM | Coffee Break

Networking opportunity with speakers and participants

10:45 AM – 12:00 PM | Generative AI with Amazon Bedrock

Foundation Models Overview

Comprehensive comparison of available models:

  • Claude (Anthropic)

    • Strengths: Long context, reasoning, safety
    • Use cases: Complex analysis, content generation

  • Llama (Meta)

    • Strengths: Open-source, customizable
    • Use cases: Fine-tuning, specialized applications

  • Titan (Amazon)

    • Strengths: Cost-effective, AWS-native
    • Use cases: Text generation, embeddings

Prompt Engineering Techniques

  • Basic Principles

    • Clear instructions and context
    • Role definition and constraints
    • Output format specification

  • Advanced Techniques

    • Chain-of-Thought (CoT) reasoning
    • Few-shot learning examples
    • Zero-shot vs few-shot comparison
    • Prompt templates and best practices

Retrieval-Augmented Generation (RAG)

  • Architecture Overview

    • Vector databases and embeddings
    • Semantic search integration
    • Context retrieval strategies

  • Knowledge Base Integration

    • Document ingestion and processing
    • Chunking strategies
    • Metadata management
    • Query optimization

Bedrock Agents

  • Multi-step Workflows

    • Task decomposition
    • Sequential reasoning
    • Error handling and retries

  • Tool Integrations

    • API connections
    • Database queries
    • External service calls
    • Custom function execution

Guardrails

  • Safety Measures

    • Content filtering
    • Harmful content detection
    • PII redaction
    • Topic restrictions

  • Content Filtering

    • Input validation
    • Output moderation
    • Compliance enforcement

Live Demo: Building a Generative AI Chatbot

  • Step-by-step chatbot creation using Bedrock
  • RAG implementation example
  • Guardrails configuration
  • Real-time Q&A with the chatbot

12:00 PM | Lunch Break (Self-arranged)

Key Highlights

Amazon SageMaker Insights

End-to-End ML Platform

  • SageMaker provides a complete solution from data preparation to model deployment
  • Integrated MLOps capabilities reduce time-to-production
  • Built-in algorithms and frameworks support various ML use cases
  • Studio interface simplifies the ML workflow

Key Features Demonstrated:

  • Automated data labeling reduces manual effort
  • Distributed training accelerates model development
  • One-click deployment simplifies production rollout
  • Model monitoring ensures ongoing performance

Generative AI with Bedrock

Foundation Model Selection

  • Different models excel at different tasks
  • Consider factors: cost, performance, context length, safety
  • Claude excels at reasoning and long-context tasks
  • Llama offers flexibility through open-source customization
  • Titan provides cost-effective AWS-native solutions

Prompt Engineering Best Practices

  • Clear, specific instructions yield better results
  • Chain-of-Thought improves reasoning quality
  • Few-shot examples guide model behavior
  • Iterative refinement is essential

RAG Architecture Benefits

  • Grounds AI responses in factual data
  • Reduces hallucinations significantly
  • Enables domain-specific knowledge integration
  • Keeps information current without retraining

Bedrock Agents Capabilities

  • Automate complex multi-step workflows
  • Integrate with existing systems and APIs
  • Handle dynamic decision-making
  • Provide transparency in reasoning process

Guardrails Importance

  • Essential for production deployments
  • Protect against harmful content
  • Ensure compliance with regulations
  • Maintain brand safety and reputation

Key Takeaways

Technical Knowledge

Amazon SageMaker:

  • ✅ Comprehensive platform for entire ML lifecycle
  • ✅ MLOps integration accelerates deployment
  • ✅ Scalable infrastructure handles large workloads
  • ✅ Cost optimization through managed services

Amazon Bedrock:

  • ✅ Multiple foundation models for different needs
  • ✅ Prompt engineering is critical for quality outputs
  • ✅ RAG architecture enhances accuracy and relevance
  • ✅ Agents enable complex workflow automation
  • ✅ Guardrails are essential for safe production use

Practical Applications

Use Cases Identified:

  • Customer service chatbots with RAG
  • Content generation and summarization
  • Code generation and documentation
  • Data analysis and insights extraction
  • Automated workflow orchestration

Implementation Considerations:

  • Start with clear use case definition
  • Choose appropriate foundation model
  • Implement RAG for domain-specific knowledge
  • Configure guardrails before production
  • Monitor and iterate based on performance

Best Practices

ML Development:

  • Use SageMaker Studio for unified development
  • Implement MLOps from the start
  • Monitor model performance continuously
  • Plan for model retraining and updates

GenAI Development:

  • Test multiple foundation models
  • Invest time in prompt engineering
  • Implement RAG for factual accuracy
  • Use agents for complex workflows
  • Always configure guardrails

Applying to Work

Immediate Actions

Explore SageMaker:

  • Set up SageMaker Studio environment
  • Experiment with built-in algorithms
  • Practice data preparation workflows
  • Test model deployment options

Experiment with Bedrock:

  • Try different foundation models
  • Practice prompt engineering techniques
  • Build a simple RAG prototype
  • Test guardrails configuration

Short-term Goals

ML Projects:

  • Identify suitable ML use cases in current projects
  • Propose SageMaker for next ML initiative
  • Implement MLOps best practices
  • Set up model monitoring

GenAI Projects:

  • Build proof-of-concept chatbot
  • Implement RAG for knowledge base
  • Create Bedrock agent for workflow automation
  • Establish guardrails standards

Long-term Vision

Organizational Impact:

  • Evangelize AI/ML adoption
  • Establish best practices and standards
  • Build reusable components and templates
  • Create knowledge sharing culture

Skill Development:

  • Pursue AWS ML certifications
  • Deep dive into specific models
  • Master prompt engineering
  • Learn advanced RAG techniques

Event Experience

Learning Environment

Professional Setting:

  • AWS Vietnam Office provided excellent facilities
  • Well-organized schedule with appropriate breaks
  • Interactive sessions encouraged participation
  • Live demos enhanced understanding

Expert Speakers:

  • AWS specialists shared real-world insights
  • Practical examples from production systems
  • Clear explanations of complex concepts
  • Responsive to questions and discussions

Hands-on Learning

Live Demonstrations:

  • SageMaker Studio walkthrough was comprehensive
  • Bedrock chatbot demo showed practical implementation
  • Real-time problem-solving demonstrated best practices
  • Interactive elements kept engagement high

Practical Insights:

  • Learned from actual production use cases
  • Understood common pitfalls and solutions
  • Gained confidence to start own projects
  • Received guidance on next steps

Networking Opportunities

Peer Connections:

  • Met fellow AI/ML enthusiasts
  • Exchanged ideas and experiences
  • Discussed potential collaborations
  • Built professional network

AWS Community:

  • Connected with AWS specialists
  • Learned about AWS resources and support
  • Discovered community events and programs
  • Identified mentorship opportunities

Personal Growth

Technical Skills:

  • Expanded knowledge of AWS AI/ML services
  • Gained practical GenAI experience
  • Understood MLOps principles
  • Learned industry best practices

Professional Development:

  • Increased confidence in AI/ML domain
  • Identified career development paths
  • Recognized areas for further learning
  • Motivated to pursue certifications

Reflections

What Worked Well

Content Quality:

  • Well-structured agenda covered essential topics
  • Balance between theory and practice
  • Appropriate depth for target audience
  • Relevant and current information

Delivery:

  • Engaging presentation style
  • Clear explanations with examples
  • Interactive demonstrations
  • Good time management

Logistics:

  • Professional venue and setup
  • Smooth registration process
  • Adequate breaks and networking time
  • Excellent organization overall

Areas for Improvement

Suggestions:

  • More hands-on lab time would be beneficial
  • Provide pre-event materials for preparation
  • Include more advanced topics for experienced participants
  • Offer follow-up sessions or office hours

Overall Assessment

The AWS Cloud Mastery Series #1 event was an excellent introduction to AI/ML and GenAI on AWS. The combination of comprehensive content, expert speakers, live demonstrations, and networking opportunities created a valuable learning experience.

Key Benefits:

  • ✅ Solid foundation in AWS AI/ML services
  • ✅ Practical knowledge of Bedrock and SageMaker
  • ✅ Confidence to start own AI/ML projects
  • ✅ Professional network expansion
  • ✅ Clear path for continued learning

Impact: This event has significantly enhanced my understanding of AI/ML on AWS and provided practical knowledge that I can immediately apply to work projects. The exposure to both SageMaker and Bedrock has opened new possibilities for implementing AI solutions.

Event Photos

Event picture 1 Event picture 2

Conclusion

Attending the AWS Cloud Mastery Series #1 was a valuable investment in professional development. The event provided not only technical knowledge but also practical insights, networking opportunities, and motivation to pursue AI/ML initiatives.

Next Steps:

  1. Complete hands-on labs with SageMaker and Bedrock
  2. Build proof-of-concept projects
  3. Share learnings with team
  4. Attend future AWS Cloud Mastery Series events
  5. Pursue AWS ML certification

The event successfully achieved its objectives of introducing AWS AI/ML services and inspiring participants to explore and implement AI solutions in their work.