GenAI Innovation Workshop — Building AI Agentic Applications

​Build production-grade AI agents on AWS — in one morning.

​Vietnamese enterprises are moving from GenAI experimentation to deployment. This hands-on workshop shows technical builders how to ship AI agents that work alongside humans, comply with local regulations, and deliver measurable ROI — using PixelML's AgenticFlow platform on AWS infrastructure.

​Who it's for 30–50 technical builders: CTOs, VPs of Engineering, Technical Leads, Solution Architects — primarily from manufacturing and scale-up enterprises (80% startup/SME).

​Why now

• ​Global GenAI spend: $644B, +76.4% YoY (Gartner 2026)

• ​2026 is the year of AI Employees — autonomous agentic workflows replacing chatbots and POCs

• ​Vietnam: 70%+ digital adoption, Decree 356/2025/NĐ-CP driving secure-AI demand, Da Nang emerging as Central Vietnam's innovation hub

​Agenda

​08:00 – 08:45 · Welcome & Registration On-site check-in, coffee, networking.

​08:45 – 09:30 · Opening — 3 Organizations, 1 Vision

• ​AWS (15 min): Introduction to Amazon Web Services

• ​PixelML (15 min): PixelML & the AgenticFlow.ai platform

• ​oBacker (15 min): The AI-enabled back-office — how oBacker uses agentic AI internally to serve hundreds of SME clients

​09:30 – 10:00 · AWS Solution Introduction — AI Infrastructure for Agentic Systems Speaker: AWS Solutions Architect

• ​Application: Amazon Q & Kiro

• ​Builder: Build & host AI agents on Amazon Bedrock

• ​Foundation: Host your own models on Amazon EC2

​10:00 – 10:10 · Tea Break

​10:10 – 12:00 · Hands-on Workshop — Build Your First Production AI Agent Speaker: AWS Solutions Architect

​By the end of the session you'll have a production-ready, multi-tenant AI analytics assistant with MCP Gateway, 3 toolsets, Cognito-authenticated chat UI, tenant isolation, guardrails, and full observability.

• ​Module 1 — Agent & Infrastructure: Strands Agent (Model, Tools, Prompt) · AgentCore Gateway, Runtime & Memory · Chat UI with Cognito + Streaming

• ​Module 2 — Toolsets: Prebaked SQL (DB views → agent tools) · API integration (booking & transactional ops) · Custom SQL (Text-to-SQL with RAG + human approval)

• ​Module 3 — Security, Multitenancy & Governance: Cedar policies + JWT + row-level security · Guardrails (content safety, PII, SQL protection) · Observability (traces, dashboards, quality metrics)

• ​Module 4 — Summary & Next Steps

​What to bring

• ​A laptop (workshop is hands-on)

• ​Curiosity about agentic AI in production

​Organized by

​PixelML · AWS · oBacker