Agentic Search: Are LLMs replacing Decades of IR Wisdom?
AI Agents are LLM-backed modules that can use tools (MCP, OpenSearch, etc.) and reason about your…
General purpose agentic loop in 40 lines of Python
When rolling your own agent, tool calling comes with annoying plumbing (packaging up arguments from the…
Synthetic RAG evaluation
Evaluating retrieval-augmented generation (RAG) is crucial for building reliable AI systems. By understanding how to generate…
Ranking is overrated: why query understanding matters more
Context depends on search that understands what users / agents want. That’s why retrieval must go…
Reasoning boosts search relevance 15-30%
In my previous article, I hypothesized reasoning agents work best with simple search tools. In agentic…
Agents turn simple keyword search into compelling search experiences
A traditional RAG system often looks like the search we’ve built for years. An LLM calls…
A Machine Learning-Driven Approach for Optimizing Hybrid Search
How to greatly improve your hybrid search results using machine learning to dynamically optimize hybrid search parameters for each query.
Grug-brained evals
Big brain spend months building perfect quality metrics. Grug brain no trust, and just want dumb labels from coworkers 👍/👎.
RAG’s big blindspot
RAG apps have a big blindspot – using actual user engagement to drive improvement. But its a hard problem. Let’s discuss!
Deep dive: Enabling pluggable storage in the OpenSearch vector engine
In 2019, OpenSearch introduced the Vector Engine, which supports three native engines: Non-Metric Space Library (NMSLIB),…
