Agentic AI Foundations
Master the design, deployment and orchestration of autonomous AI agents capable of reasoning, planning and executing complex tasks.
Key skills
- ✓Design multi-component autonomous AI agent architectures
- ✓Implement reasoning and planning systems for LLM agents
- ✓Orchestrate agent workflows with tools, memory and feedback loops
- ✓Evaluate and optimise the performance and reliability of agentic systems
The question bank for this exam is prepared after your purchase. The exam is available no later than 48 hours after payment; you will receive an email as soon as it opens.
Is this right for you?
✓ This certification is right for you if:
- →AI and ML engineers
- →Backend and full-stack developers
- →Software and solutions architects
✗ This certification is not suitable if:
- →Complete beginners in AI: first consider an introductory course on LLMs and prompting before tackling agentic systems.
- →ML experts seeking advanced fine-tuning: this certification focuses on using agents via API, not model training. Consult our Master 801 level certifications in ML Engineering.
Prerequisites & eligibility
Open registration, technical prerequisites strongly recommended (Python, APIs, LLMs).
Purchase options
Exam voucher, LMS pathway, bundle packs — build your order
Options d'achat
Composez votre parcours — voucher d'examen, préparation LMS, packs entreprise.
Skills assessed
Agentic architecture
Design autonomous agents with ReAct, Chain-of-Thought and multi-agent orchestration.
Tool integration
Connect LLM agents to APIs, databases and external systems via function calling.
Memory & Context
Implement short/long-term memory systems and advanced contextual management.
Evaluation & Safety
Measure performance, manage hallucinations and guarantee the safety of autonomous agents.
Exam format
Curriculum
Fundamentals of agentic systems
Introduction to autonomous agents and their differentiation from classical chatbots. Architectural patterns: ReAct (Reasoning + Acting), Chain-of-Thought, Tree-of-Thought. Agent taxonomy: reflexive agents, planning agents, multi-agent systems. Essential components: perception (input processing), reasoning (decision making), action (tool execution). Feedback loops and continuous improvement. Use cases: autonomous customer support, research assistants, workflow automation.
Tool use and function calling
Principles of function calling with OpenAI Functions, Anthropic Tool Use, and open-source alternatives. Definition of JSON schemas for tools (parameters, types, descriptions). Error handling and retry logic strategies. Tool chaining and action sequencing. Integration of external APIs: REST, GraphQL, webhooks. Authentication and secret management. Timeouts and fallback mechanisms. Advanced patterns: parallel tool execution, conditional branching, dynamic tool selection.
Memory and contextual management
Memory architecture for agents: short-term (conversation buffer), long-term (vector stores). Implementation of RAG (Retrieval-Augmented Generation) for agents. Vector databases: Pinecone, Weaviate, Qdrant, Chroma. Embeddings and similarity search. Context window management and token optimisation. Conversation summarisation. State persistence (Redis, PostgreSQL). Memory retrieval strategies: recency, relevance, importance. Hybrid search (dense + sparse).
Orchestration and frameworks
LangChain Agents: AgentExecutor, agent types (zero-shot, conversational, structured). LangGraph for complex workflows: state graphs, conditional edges, cycles. AutoGPT patterns: goal decomposition, self-critique, iterative refinement. CrewAI for multi-agent systems: roles, tasks, collaboration. Microsoft's Semantic Kernel: plugins, planners. Framework comparison: trade-offs, use cases. Design patterns: supervisor pattern, hierarchical agents, swarm intelligence.
How to fund your certification
3 ways to cover the cost of your certification
OPCO — Employer funding
Ask your employer to cover the cost through your OPCO (employee training plan).
Personal — €450 incl. VAT
Secure online payment by credit card (Stripe).
Business rate: €585 excl. VAT
Business — custom quote
Teams, cohorts or private catalogue: a quote tailored to your volumes and goals.
Frequently asked questions
What is the difference with a generalist LLM certification?
This certification specifically focuses on autonomous agentic systems capable of planning, reasoning and executing complex tasks via tool use. It goes beyond simple prompting to cover end-to-end intelligent system architecture.
Which frameworks are covered in the exam?
The exam mainly covers LangChain, LangGraph, and generic orchestration concepts. Questions may cover AutoGPT, CrewAI or Semantic Kernel, but in-depth mastery of LangChain is essential. Architectural principles are framework-agnostic.
Do I need access to OpenAI/Anthropic APIs to prepare?
Recommended but not mandatory. The exam assesses your conceptual and architectural understanding. However, practical experience with at least one LLM provider (OpenAI, Anthropic, or open-source via Ollama) will greatly facilitate your preparation.
Does this certification cover model fine-tuning?
No. The focus is on using existing LLMs (via API) to build agentic systems. Fine-tuning, model training and ML engineering are covered by other PROVA certifications in the AI domain.
Your certification pathway
Before, during, after — the recommended progression
Droits du candidat
Transparence
Les critères de décision et le barème sont documentés et accessibles avant l'examen.
Droit d'appel
Toute décision peut être contestée dans les 30 jours. Examen par un tiers indépendant.
Plainte
Toute personne peut signaler un dysfonctionnement. Formulaire public accessible sans compte.
