Agentic AI

Agentic AI is one of the most important shifts in modern artificial intelligence. The useful question is no longer only whether a model can respond well, but whether a system can reason, act, remember, recover, and remain controlled over time.

Kavriq’s Agentic AI curriculum is organized into two connected series.

Animated gif/diagram showing the agentic AI cycle: Observe, Reason, Plan, Act, and Reflect and their interaction with the environment.

Agentic AI Foundations

The 12-part core curriculum for learning the building blocks of agent systems.

This series covers agent loops, cognitive architecture, planning, tools, MCP, memory, multi-agent systems, guardrails, evaluation, runtime internals, and capstone projects.

Explore Agentic AI Foundations

Engineering Agent Systems

The systems-level follow-up for understanding how agents behave in production.

This series focuses on uncertainty, execution over time, state drift, retries, state machines, bounded autonomy, tool contracts, safety, observability, and practical recommendations for reliable production design.

Explore Engineering Agent Systems

Recommended Path

Start with Agentic AI Foundations if you want the components.

Continue into Engineering Agent Systems when you want the production worldview: how those components behave together under uncertainty, time, state, failure, and control.