Duong Chi Trung

Hello, I'm Duong Chi Trung

I am a |

Get in Touch

About Me

AI & Machine Learning Researcher and Engineer with strong expertise in Deep Learning, Computer Vision, and Drug Discovery, experienced in designing end-to-end research pipelines and building production-ready AI systems. Skilled in Python, SQL, PyTorch, TensorFlow, distributed training, data preprocessing, and model optimization, supported by a solid foundation in probability, statistics, and hypothesis testing. My work spans academic research and real-world engineering — integrating rigorous experimentation, reproducibility, and systems-level thinking to develop high-impact AI solutions.

Full Name: Duong Chi Trung

Date of Birth: November 11, 2004

Gender: Male

Location: Ho Chi Minh City, Vietnam

Phone: +84 337 571 104

Email: alan.trungduong@gmail.com

GitHub: github.com/HiTranh2504

LinkedIn: linkedin.com/in/duong-chi-trung

Facebook: facebook.com/duongchitrung2504

Education

Industry University of Ho Chi Minh City (IUH)

09/2022 - Present | Major: Data Science

GPA: 3.58/4.0 | TOEIC: 695

Received scholarships for academic excellence

Experience

AI Engineer

01/2026 - Present | BlueBolt Software, Vietnam

  • Designed an end-to-end YOLO-based computer vision pipeline for real-time card corner detection in production environments. Implemented robust dataset validation and augmentation strategies to handle motion blur, glare, noise, and compression artifacts from surveillance stream processing.
  • Optimized training and inference across GPU environments (AMP, batch tuning, multi-worker data loading) to ensure stable and efficient deployment. (The practical requirement demands 100\% accuracy in real-world test scenarios under operational constraints.).

AI Research Internship

07/2025 - 11/2025 | National Chung Cheng University, Taiwan

  • Excellent project under the guidance of Professor Wen-Nung Lie. In-depth research on single-image 3D Human Mesh Reconstruction using diffusion methods to generate multi-hypothesis, focusing on high-resolution models (over 6,800 vertices) and improving mesh realism through optimizing the prediction of joint, pose, and shape parameters.
  • Proposed ScoreNet-X, a new architecture for multi-hypothesis scoring in 3D Human Mesh Reconstruction include 4 main block: Normalization, MeshTokenizer, and a GraphFormer-based backbone, and implementing a Dual-head scoring mechanism for ranking and regression tasks. I conducted extensive experiments on the H36M and 3DPW benchmarks, achieving a 3–4% improvement in MPJPE & MPVE error while reducing model size and computation by more than 80%.

Brain Research Consultants

06/2024 - 07/2025 | Worldquant Brain Co., Vietnam

  • In this position, I focus on researching and developing Alphas (models predicting equity market price movements), followed by statistical analysis to assess outcomes through hypothesis testing.
  • Employing computer programming to automate repetitive tasks and leverage the capabilities.

Publications

Multidimensional Vector Ranking Algorithm for The Group Recommendation System

2025 | International Conference on Computational Collective Intelligence (ICCCI 2025). Published by Springer CCIS.

  • Researched and proposed a novel ReLU-based multidimensional vector ranking algorithm for group recommendation systems, enabling data-driven weight inference among users without predefined preferences.

A Deep Learning Model for Drug-Target Interactions Prediction in Drug Discovery

2025 | International Symposium on Information and Communication Technology (SOICT 2025). Published by Springer CCIS.

  • Researched and proposed a deep learning model based on LSTM combined with Attention, which is capable of predicting drug-protein binding affinity directly from SMILES and amino acid sequences, outperforming CNN-based baselines such as DeepDTA on the KIBA dataset. Contributing to data-driven drug discovery and virtual screening studies.

Projects

Awards & Achievements

Skills

  • Python (Data Science, Automation, Scripting)
  • SQL
  • Machine Learning & Deep Learning (Scikit-learn, PyTorch, TensorFlow), Statistical Modeling
  • Data Visualization (Matplotlib, Seaborn, Tableau)
  • Financial & Quantitative Analysis (Alpha Design, Market Research)
  • Tools: Git, LangGraph, Langchain, Hugging Face, Google Colab

Contact Me