Generative & Agentic AI
Agentic AI & AI Agents
Design autonomous AI agents that plan, use tools, and complete real work — safely.
- Duration
- 12 weeks
- Format
- Online Live
- Level
- Advanced
- Credential
- Professional Certificate
Program overview
Agents are the frontier of applied AI: systems that decompose goals, call tools, coordinate with other agents, and recover from failure. This advanced program covers agent architectures, orchestration frameworks, evaluation, and the safety engineering that autonomous systems demand.
What you'll be able to do
- Design single- and multi-agent architectures for real tasks
- Implement tool use, planning, memory, and reflection loops
- Orchestrate multi-agent workflows with modern frameworks
- Evaluate agent reliability and contain failure modes
- Apply human-in-the-loop and safety patterns for autonomous systems
Tools & technologies
- LangGraph
- OpenAI Agents SDK
- CrewAI
- MCP (Model Context Protocol)
- FastAPI
- Redis
- Playwright
Curriculum
5 modules · 12 weeks · every module ends with reviewed applied work.
Agent Foundations
- From chatbots to agents
- ReAct, planning, and reflection patterns
- Tool schemas and function calling
- Memory: short-term, long-term, episodic
Building Reliable Single Agents
- State machines with LangGraph
- Error handling and retries
- Structured task decomposition
- Browser and code-execution tools
Multi-Agent Systems
- Orchestrator and specialist patterns
- Agent communication protocols
- MCP and tool ecosystems
- Cost and latency budgets
Evaluation & Safety
- Trajectory evaluation and benchmarks
- Sandboxing and permission systems
- Human-in-the-loop design
- Governance for autonomous systems
Capstone: Autonomous Workflow
- Design an agentic system for a real workflow
- Reliability testing
- Demo and architecture defense
Projects & capstone
You graduate with evidence, not just knowledge. Key projects include:
Research agent that plans, searches, and produces cited reports
Multi-agent customer-operations workflow with human escalation
Capstone: an autonomous system for a domain you choose
Who this program is for
- GenAI engineers going deeper into autonomy
- Architects designing AI-native workflows
- R&D engineers prototyping agentic products
Eligibility
Solid Python and prior experience building with LLM APIs (our Generative AI Engineering program or equivalent). This is our most advanced AI program.
Your instructors
Sofia Marchetti
Head of Generative & Agentic AI Programs
Sofia has been building LLM-powered products since the earliest GPT-3 APIs, including retrieval systems and multi-agent workflows for enterprise clients. She designs curriculum that keeps pace with a field that changes monthly — grounded in patterns that do not.
Career opportunities
Roles this program prepares you to pursue:
- AI Agent Engineer
- AI Systems Architect
- Automation Engineer (AI)
- Applied AI Researcher
Earn a verifiable certificate
Complete all required project work to earn the Pioneer Academy Agentic AI & AI Agents certificate — digitally issued with a unique verification link for your CV and LinkedIn.
Program FAQs
How is this different from the Generative AI program?
GenAI Engineering covers the foundations of building with LLMs. This program assumes those skills and focuses entirely on autonomy: planning, tool use, orchestration, and multi-agent design.
Which frameworks will I learn?
LangGraph and the OpenAI Agents SDK are the primary stack, with comparative work in CrewAI and MCP so you can evaluate tools independently.
Take the next step
Start your GenAI & Agents journey
The next Agentic AI & AI Agents cohort is enrolling now. Applications take under ten minutes, and seats are confirmed in order of application.