55 items
Stateless LLM Calls vs Stateful Apps
Slides / VideoChoosing the Right Pattern
Slides / VideoComposability and Modular Component Design
Slides / VideoRAG Knowledge Assistant Architecture
Slides / VideoChatbot Reference Architecture
Slides / VideoSync vs Async LLM Architectures
Slides / VideoThe LLM Gateway / Proxy Layer
Slides / VideoSeparating Business Logic from LLM Calls
Slides / VideoDeterministic vs Model-Driven Control Flow
Slides / VideoWorkflows vs Agents: When to Use Each
Slides / VideoFan-Out/Fan-In for LLM Calls
Slides / VideoThe Orchestrator-Workers Pattern
Slides / VideoThe Evaluator-Optimizer Pattern
Slides / VideoLayers of an LLM App Stack
Slides / VideoWhat LLM Application Architecture Means
Slides / VideoParallelization Runs Sequentially
Code QuizOrchestrator Missing Await
Code QuizEvaluator-Optimizer Ignores Critique
Code QuizReAct Loop Drops Observations
Code QuizReflection Returns Uncritiqued Draft
Code QuizRouter Pattern Missing Fallback
Code QuizPrompt Chaining Broken Link
Code QuizRAG Retrieve-Then-Generate Flow
Code QuizRouting Tasks To The Right Architecture
Code QuizStateless vs Stateful Session Handling
Code QuizConversation State Persistence And Reconstruction
Code QuizMulti-Agent Handoff Design
Code QuizMemory Layering Short-Term vs Long-Term
Code QuizStreaming LLM Response To Client
Code QuizProvider-Agnostic LLM Client Interface
Code QuizHuman-in-the-Loop Approval Gate
Code QuizSeparating Prompt Templates From Logic
Code QuizHandling Non-Determinism in Design
QuizRouting Pattern
QuizRAG as an Architectural Pattern
QuizOrchestrator-Worker Pattern
QuizEvaluator-Optimizer Reflection Loop
QuizSingle-Prompt vs Multi-Step Design
QuizRAG vs Fine-Tuning vs Prompting
QuizBatch vs Real-Time Processing
QuizLayers of an LLM Application
QuizWhat LLM Application Architecture Is
QuizIdempotency and Retry-Safe LLM Operations
FlashcardChoosing the Right Pattern
FlashcardHandling Non-Determinism
FlashcardLLM Gateway / Proxy Layer
FlashcardAsync and Event-Driven Pipelines
FlashcardFallback and Graceful Degradation
FlashcardAgentic vs Workflow Architectures
FlashcardEvaluator-Optimizer Loop
FlashcardSeparating Prompts from Code
FlashcardParallelization Pattern
FlashcardRouter Pattern
FlashcardOrchestrator-Worker Pattern
FlashcardArchitecture for LLM Apps
FlashcardStreaming LLM Response To Client
A streaming endpoint buffers the whole response, defeating incremental delivery.
app.get('/stream', async (req, res) => {
res.setHeader('Content-Type', 'text/event-stream');
let full = '';
for await (const chunk of llm.stream(req.query.prompt)) {
full += chunk.text;
}
res.write(`data: ${full}\n\n`);
res.end();
});What is the bug in this streaming endpoint?