Showing 241–260 of 1502 insights
TitleEpisodePublishedCategoryDomainTool TypePreview
LLM-Driven Trace InsightsEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Leverage an LLM agent in Langsmith’s Insights tab to analyze thousands of traces via natural language prompts, automatically summarizing inefficiencie...
Trace-Based Debugging WorkflowEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksDevops-
Use Langsmith’s tracing projects to run backend pipelines (like YouTube transcript download and enrichment) step-by-step, surfacing errors and perform...
User-Level Feedback LoopEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Logging, tracing, and upweighting user feedback into datasets creates a feedback loop to improve both offline and online model performance and overall...
Online Referential DatasetsEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Setting up live online referential evaluation datasets enables benchmarking real application performance against optimization targets during live depl...
Offline-Online-User EvaluationEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Tom Spencer outlines a three-domain evaluation framework—offline experiments on referential datasets, online live application optimization, and user-l...
Automated Experiment LoopEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Set up an MCP server using LangChain Hub and Claude code to programmatically generate, evaluate, modify, and rerun prompt experiments in a full automa...
Prompt Performance TrackingEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksPerformance-
Use Langsmith’s prompt hub and playground to annotate, track, and iteratively optimize prompt performance across large, multi-turn datasets.
Human-in-the-Loop TogglesEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksFrontend-
The agent builder UI uses simple on/off toggles at key workflow steps to insert human review checkpoints and customize automation flows.
Action Item ExtractionEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Have the AI extract priority-ranked action items with recommended deadlines directly from email content.
Automated Email TriageEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Use an LLM-powered agent to scan a user’s inbox, classify each email by urgency and importance, and recommend or draft actions accordingly.
Agent Notification WorkflowEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Demonstrates configuring an AI agent’s notification pipeline by selecting triggers (schedule, manual, event), choosing connectors (Slack, email, Linea...
JSON-Based Email SummariesEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Passing large email sets as JSON to an agent yields a more readable summary view when dealing with voluminous inbox data.
Modular Sub-Agent PatternEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksArchitecture-
Using sub-agents that extend a primary agent and are added to a shared toolbox enables targeted functionality like urgent reply generation within a si...
Onboarding Flow PatternEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksArchitecture-
Define a custom agent by specifying its role, instructions, triggers, tools, and output format in a unified onboarding flow to streamline agent creati...
Agent Builder VisualizationEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
LangChain’s Agent Builder provides a visual graph interface to display end-to-end data flow, triggers, tools, and processes for rapid agent prototypin...
Agent Builder WorkflowEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
The AI agent creation workflow follows a structured four-step process: define agent purpose, select tools, customize parameters through user prompts, ...
No-Code Agent BuilderEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
LangChain’s Agent Builder provides a no-code, natural language interface with pre-built triggers and UI primitives to quickly assemble custom AI agent...
Abstracted Middleware PatternEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
LangChain V1 introduces an abstraction layer over core middleware components to simplify usage while preserving low-level configurability and access t...
Context Engineering WrapperEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksAi-development-
Develop intelligent wrapper layers around raw LLM requests to handle context engineering automatically and improve consistency for end users.
Next-Step Interface PatternEp 20 - The build - Claude Skills, minimax m2, LangChain Agent Builder, Options trading MCP.11/2/2025FrameworksDatabase-
Implement a ‘next-step interface’ that first retrieves schema metadata (e.g., database schema) before running analytics queries to avoid misinterpreti...
PreviousPage 13 of 76Next