This article shows the basic principles to implement a context pruning pipeline for long-running agents, based on conversational continuity and semantic relevance.
Author Archive | Iván Palomares Carrascosa
The Statistics of Token Selection: Logits, Temperature, and Top-P Walkthrough
In this article, you will learn how logits, temperature, and top-p sampling work together to control next-token prediction in large language models.
Implementing Hybrid Semantic-Lexical Search in RAG
In this article, you will learn how to implement a hybrid search strategy for RAG systems by combining BM25 lexical search with semantic search, fused together using Reciprocal Rank Fusion.
Implementing Prompt Compression to Reduce Agentic Loop Costs
In this article, you will learn what prompt compression is, why it matters for agentic AI loops, and how to implement it practically using summarization and instruction distillation.
Implementing Permission-Gated Tool Calling in Python Agents
This article illustrates how to implement a permission-gated tool in a Python agent, resulting in a robust, cost-free interception mechanism based on a simple decorator pattern.
Implementing Statistical Guardrails for Non-Deterministic Agents
Describing and implementing two simple yet effective approaches to ensure AI agent safety: semantic drift based of cosine distance and confidence thresholding based on log-probability entropy.
Effective KV Compression with TurboQuant
In this article, you will learn how TurboQuant, a novel algorithmic suite recently launched by Google, achieves advanced compression of large language models and vector search engines with no loss of accuracy.
Text Summarization with Scikit-LLM
In this article, you will learn how to use scikit-LLM’s text summarization feature to handle large volumes of text in machine learning pipelines.
A Hands-On Guide to Testing Agents with RAGAs and G-Eval
A hands-on guide to understand how to test LLM and agent-based applications using both RAGAs and frameworks based on G-Eval, concretely, by leveraging DeepEval.
Building a ‘Human-in-the-Loop’ Approval Gate for Autonomous Agents
In this article, you will learn how to implement state-managed interruptions in LangGraph so an agent workflow can pause for human approval before resuming execution.