About the course
The shift from simple LLM chat interfaces to autonomous "Agentic" systems represents the next frontier in enterprise digital transformation. However, moving from a proof-of-concept to a production-ready system requires more than just a clever prompt - it requires a robust architectural framework.
This 4-day foundations course is designed for architects and senior engineers who need to design distributed AI agent systems that are safe, reliable, and compliant. We move beyond the hype to focus on the "Sovereign AI" stack, exploring core architectural patterns, memory management, and multi-agent orchestration. The curriculum bridges the gap between probabilistic AI outputs and deterministic enterprise requirements, with a heavy focus on industries such as Healthcare, Finance, and Supply Chain.
Customisable Delivery:
This course is the first part of our 2-week Agentic Development Programme (which also comprises the Agentic Systems Deep Dive). We are happy to tailor the delivery to your team’s existing skills, specific tool preferences, and unique compliance landscape. Instructor-led online and in-house face-to-face options are available - as part of a wider customised training programme, or as a standalone workshop, on-site at your offices or at one of many flexible meeting spaces in the UK and around the World.
-
- Define Agentic Scope: Distinguish between workflows, copilots, and autonomous agents.
- Architect Persistent Memory: Design memory tiers that balance capability with token costs.
- Implement Agentic RAG: Build advanced retrieval systems with document-level security.
- Master Multi-Agent Patterns: Deploy Orchestrator-Worker and Planner-Executor architectures.
- Engineer for Safety: Establish guardrails and human-in-the-loop escalation patterns.
-
This training programme is suitable for the following audiences, and can be further tailored to include your wider organisation where appropriate:
Enterprise Architects responsible for designing the next generation of AI infrastructure.
Senior Software Engineers & Tech Leads moving from simple LLM integration to complex agent orchestration.
AI Product Managers who need to understand the technical feasibility and safety constraints of agentic products.
Security & Compliance Officers tasked with auditing AI behavior and ensuring data residency.
-
Attendees should have some experience with these disciplines:
LLM structure: A solid understanding of how Large Language Models work (tokenization, temperature, system prompts).
System Design: Familiarity with distributed systems concepts (API design, state management, event-driven architecture).
Development Experience: Proficiency in at least one modern programming language (Python is preferred for labs, but JS/TS or Java are applicable).
No prior experience with Agent frameworks (like LangGraph) is required.
We can customise the training to match your team's experience and needs - for instance with more time and coverage of fundamentals for teams at the start of their enterprise architecture journey. Get in touch to find out more.
-
This Enterprise Agentic Systems course is available for private / custom delivery for your team - as an in-house face-to-face training programme at your location of choice, or as a series of online instructor-led workshops via MS Teams (or your own preferred platform).
-
Foundations of Agentic Logic
The Autonomy Spectrum: Assistive vs. Semi-autonomous vs. Fully autonomous systems.
Agent Memory Architecture: Context window mechanics, semantic caching, and compaction strategies.
Retrieval Architecture: Agentic RAG, hybrid search, and security trimming for sensitive document access.
Architectural Patterns & Protocols
System Design: Single-agent vs. multi-agent; Planner-executor decomposition; Event-driven agent trees.
Communication Protocols: Model Context Protocol (MCP), Agent-to-Agent (A2A), and Collaboration Protocols (ACP).
Distributed Systems & Reliability
State & Idempotency: Event sourcing for agent state and managing durable memory in failure domains.
Safety Frameworks: Prompt injection defense, tool sandboxing, and output schema enforcement.
Escalation Patterns: Designing meaningful "return-to-human" triggers.
Governance, Evaluation & Ops
Compliance & Risk: PHI/PII handling, data residency, and mapping use cases to the EU AI Act risk classes.
FinOps for Agents: Token-level cost attribution and model selection strategies.
AgentOps: Trace collection, observability (LangSmith/TruLens), and chaos engineering for agents.
Trusted by
