Blog
Notes and essays on LLM workflows, agentic systems, retrieval-augmented generation, AI robustness, Vietnamese NLP, and GeoAI.
From ReAct to AFlow: The Workflow Turn in LLM Agents
1. LLM agents are no longer just prompts
Read postIndirect Prompt Injection in LLM Applications and Agents: Threat Models, Benchmarks, and Defense Mechanisms
Indirect prompt injection (IPI) refers to attacks where malicious instructions are embedded in untrusted external content (e.g., web pages, documents, tool responses) and are later consumed by an LLM application alongside the user’s intended...
Read postAdding Geospatial Features to Your Machine Learning Models
Introduction
Read postHow Natural Language Processing Can Map Human Stories
When disaster strikes, people don’t just wait for the news - they post.
Read postWhy Geography Still Matters in Our High-Tech World
What is Geography, Really?
Read post