Artificial Intelligence Ai Engineer

Nguyen Vo Thanh Duy

******.****@*****.*** 085*******

linkedin.com/in/duynvt0203 github.com/pluto0203

Summary

A motivated AI engineer with a solid academic background and practical experience gained through multiple real-world projects. I have experience in data preprocessing, model training, and optimization, with a particular focus on computer vision. Through hands-on involvement in various AI initiatives, I developed strong skills in building and deploying end-to-end pipelines. I am looking for an opportunity to join a team where I can continue learning, take on real challenges, and grow into a strong AI engineer. Technical Skills

Programming Languages: Python

Frameworks: TensorFlow, PyTorch, OpenCV, YOLO, LangChain, HuggingFace, SQL, ... Hardware & Deployment: Raspberry Pi

Tools: Git, Docker, LabelImg, CVAT, Jira

Education

FPT University, B.Sc. in Artificial Intelligence Oct 2021 – Jun 2025

• GPA: 3.2/4.0

• First Prize – "Philosophy Talk 13", 2025

• Impression Award - "FPT University Dormitory Review", 2024

• Top 5 teams in the FPT University’s Startup program, 2024 Experience

AI Engineer Intern, FPT Quy Nhon AI Center Jan 2024 – May 2024

• Built and preprocessed datasets for abnormal behavior detection in the banking domain, optimizing data quality to improve model accuracy.

• Designed and implemented a training pipeline for intruder detection, deployed on Raspberry Pi 5, enabling real-time recognition with latency under 10 seconds and meeting demo requirements.

• Participated in image data labeling using professional tools, ensuring input quality with over 95% labeling accuracy, which helped shorten project completion time. Projects

Drunk Detection System - Implement in car GitHub Repo

• Developed a real-time drunk detection system using Raspberry Pi, integrating a camera for facial analysis and an MQ-3 alcohol sensor for enhanced accuracy.

• Built a facial dataset of Asian people exhibiting visual signs of intoxication, specifically tailored for drunk detection tasks to improve the system accuracy.

• Integrated Arduino to read alcohol levels and send alerts when both AI prediction and sensor data confirmed drunk behavior.

• Developed a dashboard that allows users to efficiently manage and access history, improving usability and system transparency.

Top 10 – FPT Edu Digital Race 2023 (Group C)

• Collaborated in a 4-member team to build a dataset for traffic sign recognition module for an autonomous vehicle system.

• Collected and processed image data, applying various augmentation techniques to enhance dataset diversity and improve model accuracy.

• Achieved Top 10 ranking nationwide among Group C participants. RAG Chatbot for Vietnamese Law Documents GitHub Repo

• Built a RAG chatbot to answer questions about Vietnamese law documents using LangChain.

• Designed a full pipeline: document chunking, embedding, vector storage, and LLM-based generation for context-aware responses.

• Developed with Streamlit UI with query input, chat history, and configurable retrieval settings. Traffic Signs Recognition GitHub Repo

• Built a custom dataset of six commonly used traffic sign classes, including image collection, annotation, and preprocessing, similar to realistic environments.

• Developed classification model using YOLOv8, fine-tuned for real-time accuracy and deployed for consistent performance.

• Built a web-based demo application allowing users to test traffic sign recognition through image upload, showcasing model predictions and bounding boxes interactively. Languages

English: Upper Intermediate (comfortable in technical documentation) Japanese: Beginner

Contact this candidate