An Hung Nguyen

COMPUTER SCIENTIST / AI RESEARCHER

Danang, Vietnam

Personal Profile

AI Engineer with 5+ years of experience in Computer Vision, bridging academic research and real-world applications. Experienced in building end-to-end machine learning pipelines, from data analysis and model training to optimization and deployment on edge devices. I also have experience working with Large Language Models (LLMs) and helping customers improve the reliability, scalability, and cost-efficiency of AI solutions.

I have authored and co-authored 10+ research papers and received several academic awards throughout high school and university. Passionate about research and continuous learning, I enjoy exploring new technologies, tackling challenging problems, and turning innovative ideas into practical solutions.

Education

University of Science and Technology - The University of Danang

Danang, VN
  • Grade: 3.65/4.0
  • Excellent Bachelor's Degree
  • Working as an AI Researcher with Minh Tuan Pham PhD.
  • Capstone Project: A Deep Learning-Based Approach for Multi-Weather Image Restoration and Enhancement. (http://thuvienso.dut.udn.vn/handle/DUT/16775)
  • Courses: Data Science, Programming for Data Science, Knowledge Representation and Reasoning, Statistical Theory and Methods, Learning Skills through Case Studies, Artificial Intelligence, Machine Learning, Statistical Learning

Le Quy Don upper-secondary school for the gifted

Quang Tri, VN
  • Specialised in Biology, Informatics, and Mathematics with Computer Science
  • Selected as a member of the national-level scientific research competition team and the Informatics gifted student team.

Work Experience

SLA Software Logistik Artland GmbH

Germany (Remote)
  • Developed and deployed automation systems for low-code AI model training and data management as part of the Dubai-Premiere: TrAInity – AI-Empowerment project.
  • Built an end-to-end automated pipeline for image classification and object detection, covering dataset analysis (EDA), preprocessing, model training, and evaluation. Simplified the workflow so users only need to provide datasets and select the number of GPUs, with automated training reports generated upon completion.

SLA Software Logistik Artland GmbH

Germany (Remote)
  • Developed an OCR module for real-time recognition and tracking of slaughter numbers printed on animal hides, supporting automated traceability (One of them: https://lnkd.in/p/gufFDTc6).
  • Supported the integration and deployment of AI models into existing customer infrastructures, including hardware selection, automated installation scripts, and system monitoring and troubleshooting.
  • Developed end-to-end conversion and optimization scripts to convert PyTorch and TensorFlow models into NPU-compatible formats for efficient inference on edge devices.
  • Collaborated with an AI team to develop neural network architectures to detect and segment animal pathogens.
  • Developed scripts and applications to automate and handle data.
  • Automated and optimized the data handling process for signal processing, worked with Linux distributions, performed shell scripting, and utilized various Linux tools.
  • Technical Skills: Python with PyTorch, NumPy, Matplotlib, Pandas, Scikit-learn, C++, Ubuntu Linux, Linux tools, Apt, Scripting, Git.
  • Soft Skills: Teamwork, Time Management, Communication, Presentation skills.
  • Research Skills: Medical Research, Data Analysis, Data Engineering.

University of Regina

Canada
  • Researched solutions for simulating chemical processes.
  • Implemented simulation code using Matlab and Python.

Phygital Labs

Vietnam
  • Worked as an AI intern on deep learning projects.
  • Worked with LLM models to propose solutions for querying large documents.
  • Collaborated with a former Senior Google engineer to deploy systems.

FX Data

Germany (Remote)
  • Collaborated with an AI team to develop neural network architectures to detect and segment animal pathogens.
  • Developed scripts and applications to automate and handle data.
  • Automated and optimized the data handling process for signal processing, worked with Linux distributions, performed shell scripting, and utilized various Linux tools.
  • Technical Skills: Python with PyTorch, NumPy, Matplotlib, Pandas, Scikit-learn, C++, Ubuntu Linux, Linux tools, Apt, Scripting, Git.
  • Soft Skills: Teamwork, Time Management, Communication, Presentation skills.
  • Research Skills: Medical Research, Data Analysis, Data Engineering.

Rhen Consulting

Germany (Remote)
  • Collaborated with an AI team to develop neural network architectures to detect and segment animal pathogens.
  • Developed scripts and applications to automate and handle data.
  • Automated and optimized the data handling process for signal processing, worked with Linux distributions, performed shell scripting, and utilized various Linux tools.

Highlight Awards

Fourth Place Award in the Field of Science, Engineering, and Technology

Vietnam

Project title: Quaternion Convolutional Neural Networks for Image Depth Estimation

Fourth Place Award in the in Computational Biology and Bioinformatics

Vietnam

Project title: Applying Deep Learning to the process of diagnosing, analyzing and detecting damage to the retina due to diabetes

Third Prize in the Provincial Contest for Excellent Students in Informatics

Vietnam

Publications

All-in-One Image Restoration with Quaternion Cross-Attention and Haar Frequency Decomposition

Optimizing Retrieval in Vietnamese Literary QA Using Named-Entity Relationship Graphs

  • Status: Under review
  • Minh Tuan Pham, An Hung Nguyen, Phuc Hao Do, Thanh Huy Nguyen, and Nam Do

Real-Time 3D Reconstruction through Image Synthesis and IoT Integration

  • Cong Thang PHAM, An Hung Nguyen, Tan Sang Truong, Dai Quy Tran, Si Tien Tran
  • Submitted to Vestnik St. Petersburg University: Mathematics

Adaptive Hybrid Transformer Network with Frequency Attention for All-in-One Image Restoration

DehazeCLNet: A Contrastive Learning Framework with Advanced Feature Extraction for Image Dehazing

Enhancing Image Classification Capabilities In The Vision Transformer Network Model With Quaternion Algebra

Enhancing Convolutional Neural Network Architectures with Long Short-Term Memory for Improved Image Classification

Quaternion Convolutional Neural Networks for Depth Estimation

Weight Initialization for Complex Number Convolutional Neural Networks Using Genetic Algorithms

Efficient Complex Valued Neural Network with Fourier Transform on Image Denoising

Skills

Top honors and awards

Languages