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
- 01 What an agent actually is
- 02 The agent harness pattern
- 03 Tool calling and function design
- 04 Planning and task decomposition
- 05 Memory architectures, short, long and episodic
- 06 Self-correction and reflection loops
- 07 Multi-agent orchestration and where it fails
- 08 Sandboxing and safe code execution
- 09 Human oversight and interruption
- 10 Evaluating agent behaviour
- 11 Cost control and token economics
- 12 Failure modes and debugging
- 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.