PinnedNgoc·Jun 30Build AI Agent From Scratch in PythonHow to build an AI Agent from scratch in Python, include tool calling, dependency injection, structured outputs, and complete agent loop
PinnedNgoc·Jul 6, 2025Why Semantic Parsing is So Painful — My GraphRAG JourneyGraphRAG (Graph Retrieval-Augmented Generation) holds a lot of promise, but making it work in practice is painful — from building the…A response icon8A response icon8
PinnedNgoc·Jan 14, 2025Part 1: How to Effectively Build a Document-centric Graph for History Question Answering from TextIn this post, I walk you through how to construct a document-centric graph for history question answering
Ngoc·Jun 13, 2025From Human-in-The-loop Knowledge graph Construction to Agentic GraphRAG.A practical guide to building robust, purpose-driven knowledge graphs in the medical domain using human-in-the-loop agentic workflows…
Ngoc·May 27, 2025Build an Agent for Document-Based Knowledge Graphs Construction (GraphRAG)This post introduces a modular, agent-based approach to building document-centric knowledge graphs using PydanticAI. By delegating tasks…A response icon1A response icon1
Ngoc·May 25, 2025Learn to Deploy LLM Apps: Run Ollama on Google Colab and Connect Locally with NgrokBuilding an LLM project directly in a notebook can be challenging, especially when it comes to modularity and management. In this post…
Ngoc·May 24, 2025How I Run Ollama on Google Colab to Learn AI Agents with Pydantic AIRun powerful local LLMs in Google Colab using Ollama, and build structured AI agents with Pydantic AI…
Ngoc·May 21, 2025Part 3A: This is how I use PyG to load and save Document-based GraphIn this post, I’ll walk you through how I use PyTorch Geometric (PyG) to build and work with a document-based heterogeneous graph…
Ngoc·Jan 22, 2025Part 2A: Implementing a Graph Builder to Extract a Basic Document-Centric Knowledge Graph in Python.In this post, I’ll show you how to construct a basic document-centric knowledge graph using Python and Local LLM
Ngoc·Jan 15, 2025How to eliminate the chance of generating Chinese in Qwen 2.5Qwen 2.5 is a powerful language model known for its strong performance in text generation and complex tasks. Compared to other LLMs like…