Hello and welcome! I’m Phong Do — an AI researcher passionate about how machines understand language, and how that understanding connects with the world around us.

My journey into AI began at the NLP@UIT lab at University of Information Technology in Ho Chi Minh City, where I first explored Natural Language Processing as an undergraduate student. This early exposure laid the foundation for a lasting commitment: I’ve continued my research with the lab even after graduation. In parallel, I’ve worked at Zalo AI for over three years, focusing on both speech processing and NLP, with a particular emphasis on large language models. At Zalo, I contributed to the development of speech processing models for Kiki Assistant, a Vietnamese virtual voice assistant installed in over 1 million cars, helping bring conversational AI into daily driving experiences.

But beyond code and computation, there’s another part of me: the geographer at heart. I’ve always been captivated by the way places shape people, and how language reflects the landscapes we come from. This passion drives me to explore not only AI and Geography as separate disciplines, but also the rich, underexplored intersection between them — where spatial thinking can inform smarter, more human-centered AI.

Whether through research, writing, or collaboration, I’m on a path to build systems that understand more than just words — systems that understand context, culture, and connection.

Research Interests

I’m passionate about building intelligent retrieval systems that can reason, generalize across modalities and languages, and interact with structured knowledge. My research focuses on two major directions: enhancing Retrieval-Augmented Generation (RAG), and exploring the intersection between Artificial Intelligence and Geography.

• 🔍 Retrieval-Augmented Generation: I’m interested in advancing RAG systems to make retrieval more intelligent, interpretable, and adaptable to complex data types.

  • 📘 Knowledge Graph Reasoning: Leveraging structured knowledge graphs to support logical reasoning and evidence chaining during retrieval.

  • 🌐 Multilingual & Multimodal Embedding: Designing unified embedding spaces for cross-language and cross-modal semantic search - enabling systems to search across text, images, and more.

• 🗺️ AI and Geography: As a geography enthusiast, exploring the intersection between spatial intelligence and artificial intelligence.

  • 🧭 Spatial RAG & Geographic Question Answering: Developing systems that understand and retrieve geospatial information, enabling models to answer geography-based questions and reason over maps, locations, and spatial relationships.

I am happy to chat and discuss potential collaborations. Please feel free to reach out to me via Email (phongdntvn@gmail.com).

📰 News