Building LLM Applications, RAG & AI Agents4 Sections28 Lessons4 Days
Curriculum
- 4 Sections
- 28 Lessons
- 4 Days
Expand all sectionsCollapse all sections
- Day 1: LLM application architecture7 Items
- 1.1LLM application architecture.
- 1.2Compare model, application and orchestration responsibilities
- 1.3Define latency and quality requirements
- 1.4Inputs, records and relevant assumptions for llm application architecture.
- 1.5Applied exercise, feedback and review of common misunderstandings
- 1.6Summary of learning priorities and connection to the wider programme.
- 1.7Practical workshop: Sketch a bounded architecture.
- Day 2: Embeddings, vector search and RAG and Tool use and structured outputs7 Items
- 2.1Embeddings, vector search and RAG.
- 2.2Chunk documents, build retrieval and examine relevance
- 2.3Related requirements, data inputs and practical judgements for embeddings, vector search and rag
- 2.4Tool use and structured outputs.
- 2.5Validate structured outputs and tool arguments
- 2.6Restrict permissions
- 2.7Practical workshop: Complete linked exercises: compare retrieval settings on a supplied corpus; implement a controlled tool call.
- Day 3: Agentic workflow patterns and Evaluation, observability and guardrails7 Items
- 3.1Agentic workflow patterns.
- 3.2Design explicit workflow states, stopping conditions and human approvals
- 3.3Related requirements, data inputs and practical judgements for agentic workflow patterns
- 3.4Evaluation, observability and guardrails.
- 3.5Create evaluation cases for groundedness, prompt injection and tool misuse
- 3.6Related requirements, data inputs and practical judgements for evaluation, observability and guardrails
- 3.7Practical workshop: Complete linked exercises: build a limited multi-step agent; record failures and corrective controls.
- Day 4: Deployment, cost and governance7 Items
- 4.1Deployment, cost and governance.
- 4.2Estimate token cost and latency
- 4.3Plan versioning, rollback and monitoring
- 4.4Inputs, records and relevant assumptions for deployment, cost and governance.
- 4.5Applied exercise, feedback and review of common misunderstandings.
- 4.6Summary of learning priorities and connection to the wider programme.
- 4.7Practical workshop: Present a deployable prototype and operational checklist.

