LLM Application Architecture & Design Patterns

55 lessons in AI Engineering / Working with LLMs

• Stateless LLM Calls vs Stateful AppsSlides / Video
• Choosing the Right PatternSlides / Video
• Composability and Modular Component DesignSlides / Video
• RAG Knowledge Assistant ArchitectureSlides / Video
• Chatbot Reference ArchitectureSlides / Video
• Sync vs Async LLM ArchitecturesSlides / Video
• The LLM Gateway / Proxy LayerSlides / Video
• Separating Business Logic from LLM CallsSlides / Video
• Deterministic vs Model-Driven Control FlowSlides / Video
• Workflows vs Agents: When to Use EachSlides / Video
• Fan-Out/Fan-In for LLM CallsSlides / Video
• The Orchestrator-Workers PatternSlides / Video
• The Evaluator-Optimizer PatternSlides / Video
• Layers of an LLM App StackSlides / Video
• What LLM Application Architecture MeansSlides / Video
• Parallelization Runs SequentiallyCode Quiz
• Orchestrator Missing AwaitCode Quiz
• Evaluator-Optimizer Ignores CritiqueCode Quiz
• ReAct Loop Drops ObservationsCode Quiz
• Reflection Returns Uncritiqued DraftCode Quiz
• Router Pattern Missing FallbackCode Quiz
• Prompt Chaining Broken LinkCode Quiz
• RAG Retrieve-Then-Generate FlowCode Quiz
• Routing Tasks To The Right ArchitectureCode Quiz
• Stateless vs Stateful Session HandlingCode Quiz
• Conversation State Persistence And ReconstructionCode Quiz
• Multi-Agent Handoff DesignCode Quiz
• Memory Layering Short-Term vs Long-TermCode Quiz
• Streaming LLM Response To ClientCode Quiz
• Provider-Agnostic LLM Client InterfaceCode Quiz
• Human-in-the-Loop Approval GateCode Quiz
• Separating Prompt Templates From LogicCode Quiz
• Handling Non-Determinism in DesignQuiz
• Routing PatternQuiz
• RAG as an Architectural PatternQuiz
• Orchestrator-Worker PatternQuiz
• Evaluator-Optimizer Reflection LoopQuiz
• Single-Prompt vs Multi-Step DesignQuiz
• RAG vs Fine-Tuning vs PromptingQuiz
• Batch vs Real-Time ProcessingQuiz
• Layers of an LLM ApplicationQuiz
• What LLM Application Architecture IsQuiz
• Idempotency and Retry-Safe LLM OperationsFlashcard
• Choosing the Right PatternFlashcard
• Handling Non-DeterminismFlashcard
• LLM Gateway / Proxy LayerFlashcard
• Async and Event-Driven PipelinesFlashcard
• Fallback and Graceful DegradationFlashcard
• Agentic vs Workflow ArchitecturesFlashcard
• Evaluator-Optimizer LoopFlashcard
• Separating Prompts from CodeFlashcard
• Parallelization PatternFlashcard
• Router PatternFlashcard
• Orchestrator-Worker PatternFlashcard
• Architecture for LLM AppsFlashcard