Cisco AI Technical Practitioner Bootcamp

• LLMs, diffusion and multimodal models [1.1]
• Summarization, content creation and code generation [1.1]
• How generative models produce output
• Where AI fits in IT operations

LAB Match use cases to model families

• Cloud-hosted vs. locally hosted models [1.2]
• Cost, latency, privacy and scalability trade-offs [1.2]
• Choosing models in hubs and repositories [1.4]
• Reasoning and multimodal capabilities [1.4]

LAB Select and justify a model for three scenarios

• Tokens and token limits [1.3]
• Context windows [1.3]
• Managing long inputs and responses [1.3]
• Token cost and latency in prototypes [5.4]

LAB Estimate tokens and fit a task in context

• Retrieval augmented generation [1.5]
• Embeddings and similarity search [1.5]
• Vector databases [1.5]
• When to use RAG vs. fine-tuning

LAB Design a RAG pipeline for an IT knowledge base

• Roles, instructions and constraints [2.1]
• Prompt patterns [2.1]
• Structuring prompts for text output [2.2]
• Evaluating and iterating on output

LAB Rewrite weak prompts into reliable ones

• Iterative and sequential prompting [2.2]
• Chained prompts [2.2]
• Few-shot prompting [2.2]
• Prompting for image and audio generation [2.2]

LAB Build a prompt chain for incident summaries

• Prompt injection attack types [2.3]
• Hallucinations and AI-generated errors [2.4]
• Defensive prompting [2.4]
• Mitigation strategies [2.4]

LAB Detect and defend against prompt injection

• Fairness, transparency, accountability and safety [3.1]
• Protecting corporate data in AI systems [3.2]
• AI-specific threats and misinformation [3.3]
• Policy, risk management and compliance [3.4]

LAB Write an AI acceptable-use policy

• AI in exploratory data analysis [4.1]
• Automated data preparation [4.2]
• Privacy controls for AI-assisted analysis [4.3]
• AI-assisted research and drafting [4.4]

LAB Clean and analyze a dataset with AI assistance

• AI from requirements to deployment [5.1]
• Code generation and rapid prototyping [5.2]
• Debugging, error handling and documentation [5.5]
• Reviewing AI-written code safely

LAB Prototype and debug a script with AI

• Workflow design principles [5.3]
• Monitoring AI workflows [5.3]
• Token usage, cost and latency [5.4]
• Measuring output quality

LAB Design a monitored AI workflow

• Agentic vs. generative AI [6.1]
• Agent design, autonomy and orchestration [6.2]
• Model Context Protocol primitives [6.3]
• Human-in-the-loop and data mapping [6.4-6.5]

LAB Capstone: design an IT operations agent