Nguyễn Công Tiến Nguyễn Công Tiến

I’m an AI Engineer with hands-on experience in Vietnamese NLP, LLM fine-tuning, RAG, document AI, OCR, semantic search, and MLOps. I build systems that go from data preparation and evaluation to deployment-ready pipelines.

In recent work at AILOGY, I built and maintained production NLP/LLM services using Flask, Docker, and Ubuntu servers. I also designed MLOps workflows for data pipelines, automated evaluation and reporting, and experiment tracking with Weights & Biases.

I focus heavily on retrieval and end-to-end quality: I’ve implemented training, evaluation, indexing, and deployment pipelines using tools like FAISS, Qdrant, and Elasticsearch. I care about measurable performance using Accuracy, macro-F1, BLEU, and ROUGE—then validating with qualitative analysis.

I also enjoy research and practical engineering in parallel, from Vietnamese instruction-tuning and prompting strategies to document intelligence pipelines for PDFs and scanned images. Whether it’s packaging models with ONNX or building hybrid search with metadata-aware ranking, I aim for reliable, fast, and maintainable AI systems.