Tool-Calling Design Patterns for Search-Grounded LLM Agents
Architectural patterns determine whether search-grounded agents retrieve and reason reliably.
Linh Phuong Dao
LLM Evals Weekly — Page 2
Agent Search
Web Search APIs Built for AI Agents and LLMs
AI agents need search APIs built for machines, not browsers.
Multi-Step Query Refinement in Agentic Research Workflows
Iterative verification catches errors that single-pass research agents miss.
Camille Roth-Svensson
More From LLM Evals Weekly
Real-Time Web Data Providers for LLM Knowledge Cutoff Problems
Retrieval layers now separate production systems from hallucinations about current events.
Marcus Delacroix-Okafor
Web Search API Comparison for LangChain and LlamaIndex Agents
LlamaIndex handles retrieval, LangGraph handles orchestration, each needs different search APIs.
Camille Roth-Svensson
RAG Evaluation Frameworks for Grounded Answer Quality
Split retrieval and generation scoring to catch hallucinations end-to-end evaluation misses.
Linh Phuong Dao
Structured Data Extraction From Web Pages for RAG Grounding
A four-stage pipeline ensures RAG data stays grounded instead of fabricated.
Tariq Osei-Mensah
Chunking Strategies for Live Web Content in RAG Pipelines
Query-time fetching forces a rethink of how web pages get split.
Linh Phuong Dao
Hybrid Retrieval Combining Dense Vectors and Live Web Search
Dense vectors miss SKUs and keywords miss paraphrases, so fuse both.
Tariq Osei-Mensah
Freshness vs. Relevance Trade-Offs in Document Retrieval
Semantic search can't tell old from new, and that gets costly.
Linh Phuong Dao
Retrieval Latency Budgeting in Production RAG Systems
Fixed time limits at each pipeline stage stop timeout cascades.
Tariq Osei-Mensah
Citation and Source Attribution Pipelines for LLM Responses
Averaging citation scores hides failures, so score each rubric separately.
Camille Roth-Svensson