Advanced AI Agents & Production Engineering

A 3-day advanced intensive for AI Accelerator graduates — multi-agent systems, A2A interoperability, advanced MCP, security hardening, and enterprise operations

Course Format and Delivery

Delivery Method: LiveOnline
Schedule: 3 days

All sessions are delivered live by an expert instructor in a fully interactive online environment.

*20% off for group bookings when booking 3 or more attendees from the same organization on the same course dates in the same transaction.

About this course

This 3-day intensive is the 'Level 2' follow-on to the Enterprise AI Accelerator. Where the Accelerator builds foundational fluency across the AI stack, this program assumes that foundation and goes deep on production engineering: multi-agent orchestration with LangGraph, CrewAI, AutoGen, and the Claude Agent SDK; agent-to-agent interoperability via Google's A2A protocol (v1.0, 150+ organizations); complex MCP server architectures; security hardening drawing from the Enterprise AI Security program's six-layer blueprint (agent hijacking, MCP auth/scoping, supply-chain defense); full-stack observability with LangSmith; cost optimization with model routing and token budgets; and a team capstone building a deployable enterprise multi-agent system with A2A connectivity.

What you will need

Completion of the Enterprise AI Accelerator or equivalent experience with agents and MCP. Comfort with Python, LLM APIs, basic agent patterns, and MCP concepts. This is not an introductory program.

Consolidates and extends advanced content from the Production Agents (#8), Multi-Agent (#9), Observability (#11), and Security workshops into a cohesive 3-day journey. Approximately 30% lecture, 70% hands-on labs

Topics Covered

Day 1: Advanced Agent Orchestration & Multi-Agent Systems

Time	Topic	Type
9:00 – 9:30	Welcome, prerequisites check, environment setup	Setup
9:30 – 10:15	Beyond ReAct: plan-and-execute, reflection, tool-learning architectures	Lecture
10:15 – 11:00	Lab: Build a self-reflecting agent with dynamic tool selection	Hands-on
11:00 – 11:15	Break	Break
11:15 – 12:00	Multi-agent patterns: supervisor, hierarchical, collaborative, swarm	Lecture
12:00 – 12:45	Lab: Build a supervisor multi-agent system with LangGraph + CrewAI	Hands-on
12:45 – 1:45	Lunch	Break
1:45 – 2:30	A2A protocol: Agent Cards, task lifecycle, gRPC transport, discovery	Lecture
2:30 – 3:30	Lab: Build A2A-compliant agents and connect them across frameworks	Hands-on
3:30 – 3:45	Break	Break
3:45 – 4:30	Claude Agent SDK + Google ADK: building custom multi-agent systems	Lecture
4:30 – 5:15	Lab: Build a production multi-agent system with A2A interop	Hands-on
5:15 – 5:30	Day 1 recap and Day 2 preview	Discussion

Day 2: Production Hardening, MCP & Security

Time	Topic	Type
9:00 – 9:15	Day 1 review and questions	Discussion
9:15 – 10:00	Advanced MCP: multi-server architectures, dynamic registration, transport optimization	Lecture
10:00 – 10:45	Lab: Build a multi-server MCP architecture with dynamic discovery	Hands-on
10:45 – 11:00	Break	Break
11:00 – 11:45	Agent security: goal hijacking, data exfiltration, Semantic Kernel RCE case study	Lecture
11:45 – 12:30	Lab: Attack and defend — exploit and patch agent vulnerabilities	Hands-on
12:30 – 1:30	Lunch	Break
1:30 – 2:15	MCP security: JWT auth, per-tool scopes, sandboxing, supply chain defense	Lecture
2:15 – 3:00	Lab: Implement auth, scopes, rate limiting, and defense-in-depth for MCP	Hands-on
3:00 – 3:15	Break	Break
3:15 – 4:00	Production resilience: error budgets, circuit breakers, fallback chains	Lecture
4:00 – 4:45	Lab: Add resilience patterns with fallback and circuit-breaker logic	Hands-on
4:45 – 5:15	Human-in-the-loop: approval workflows, confidence-based routing	Lecture
5:15 – 5:30	Day 2 recap and capstone introduction	Discussion

Day 3: Observability, Cost Optimization & Capstone

Time	Topic	Type
9:00 – 9:15	Day 2 review and capstone team formation	Discussion
9:15 – 10:00	Full-stack observability: LangSmith tracing, eval suites, cost tracking	Lecture
10:00 – 10:45	Lab: Instrument a multi-agent system with comprehensive observability	Hands-on
10:45 – 11:00	Break	Break
11:00 – 11:45	Cost optimization: model routing/cascading, token budgets, FinOps	Lecture
11:45 – 12:30	Lab: Build cost-optimized agents with cascading and budget controls	Hands-on
12:30 – 1:30	Lunch	Break
1:30 – 3:30	Capstone: Build a production multi-agent enterprise system with A2A + MCP	Hands-on
3:30 – 3:45	Break	Break
3:45 – 4:30	Capstone presentations and peer review	Discussion
4:30 – 5:00	Enterprise deployment playbook, scaling patterns, wrap-up	Lecture

Brent Laster

Facilitator

Brent Laster is a global trainer, author, speaker, and founder/president of Tech Skills Transformations LLC. He helps enterprise teams adopt modern software practices in AI engineering, AI-assisted development, DevOps, automation, and secure software delivery. He is the author of Learning GitHub Copilot, Learning GitHub Actions, Professional Git, and Jenkins 2: Up and Running, as well as multiple online and live training programs for companies such as O'Reilly. In addition to AI expertise, Brent brings more than 25 years of experience in software development, management and technical leadership, DevOps, release engineering, and open-source technologies. He regularly presents and conducts workshops at industry conferences and for private clients.