AI Automation - Course 1

Course Description

This course moves beyond individual AI prompts into the world of AI agents and automated workflows — systems that can take actions, use tools, manage files, and complete multi-step tasks on your behalf. Learners will progress from no-code automation tools through to structured agent frameworks used in industry, gaining transferable skills applicable to any organization or platform.

The course is structured around a layered tool strategy: early modules introduce accessible no-code and low- code tools (Claude Cowork, n8n) to build automation intuition, while later modules introduce the leading open- source agent frameworks — LangChain, LangGraph, and CrewAI — through guided Jupyter Notebook labs on Google Colab. Throughout, the emphasis remains practical and workforce-relevant.

The course closes with a dedicated capstone module on responsible agentic AI, grounded in published frameworks including the NIST AI Risk Management Framework (AI RMF 1.0), covering how to deploy agents safely, audit their behavior, and communicate automation policies to institutional stakeholders.

Course Learning Outcomes

Upon successful completion of this course, learners will be able to:

1. Design a comprehensive AI agent system specification — selecting and justifying the reasoning architecture, prompt engineering strategy, memory configuration, retrieval pipeline, and multi-agent coordination pattern — that coherently addresses a defined task structure, user population, and deployment context, and demonstrates the architectural reasoning and computational thinking required by the course.
2. Implement functional, end-to-end AI agent pipelines — integrating tool registries, vector embedding and retrieval systems, conversational memory modules, stateful multi-agent workflows, and production observability instrumentation — using current industry frameworks, with demonstrated correctness, modularity, and reproducibility.
3. Diagnose failures in AI agent systems by analyzing reasoning traces, RAGAS evaluation outputs, multi-agent coordination logs, and observability dashboards — attributing each observed deficiency to a specific architectural cause rather than a surface symptom and proposing structurally grounded remediations with predicted behavioral outcomes.
4. Evaluate an AI agent system’s production readiness against a multidimensional framework — spanning task performance accuracy, tool and retrieval reliability, coordination efficiency, token cost, latency, and safety behavior — formulating evidence-based architectural recommendations and comparing multi-agent designs against single-agent baselines to determine when added complexity is warranted.
5. Apply responsible AI governance frameworks — including the EU Artificial Intelligence Act risk classification tiers and the NIST AI Risk Management Framework — to produce the full set of accountability and compliance documentation required for a production agent deployment: risk register, system card, incident response plan, and data provenance statement, communicated with precision to both technical and non-technical organizational stakeholders.