Advanced engineering Engineering Available on request

Agentic AI Engineering

Most of what is marketed as an AI agent is a chatbot with extra buttons. This is how the real ones are built.

Tool calling, planning, memory and orchestration, with the cost control and human oversight that keep an autonomous system accountable.

Duration
7.5 days, 45 contact hours
Cohort size
6 to 12 participants
Delivery
In-house, Online, Blended
Languages
Arabic, English

Who it is for

Senior engineers, AI architects, technical leads building autonomous systems.

What people leave able to do

  • Design an agent architecture appropriate to a real task, not a demo
  • Build tool-calling systems with reliable error recovery
  • Implement planning, memory and self-correction loops
  • Orchestrate multiple agents without creating an unaccountable mess
  • Evaluate, monitor and control the cost of agentic systems in production

Modules

  1. 01 What an agent actually is
  2. 02 The agent harness pattern
  3. 03 Tool calling and function design
  4. 04 Planning and task decomposition
  5. 05 Memory architectures, short, long and episodic
  6. 06 Self-correction and reflection loops
  7. 07 Multi-agent orchestration and where it fails
  8. 08 Sandboxing and safe code execution
  9. 09 Human oversight and interruption
  10. 10 Evaluating agent behaviour
  11. 11 Cost control and token economics
  12. 12 Failure modes and debugging
  13. 13 Capstone: a working autonomous agent for a real task

Prerequisites

Strong Python, API experience, and ideally LLM Foundations first.

Start with a short call

Fifteen minutes to understand what your team does and what you want to change. If training is not the right answer, I will say so.