Lê Trọng Phúc
AI engineer - Data analyst
Dob: **/05/2004
Gender: Male
Phone: 091*******
Email: **********@*****.***
Address: *** ***** ** *, **** Tri Dong, Ho Chi Minh city OBJECTIVE
I am currently a final-year student majoring in Computer Science. Seeking a position as an AI Engineer or Data Analyst where I can apply my knowledge of artificial intelligence, data analysis, and programming skills to develop intelligent solutions and deliver actionable insights that support business decisions. EDUCATION
2022 - 2026 HCMC University of Foreign Languages and Information Technology - HUFLIT Data science
Current GPA: 3.12/4
PROJECT
5/2025 - 8/2025 DUCK-Net with Attention
Developed a custom CNN model (DuckNet with Attention) for semantic segmentation of gastrointestinal images.
Utilized the Kvasir dataset to train the model on polyp segmentation tasks. Implemented in PyTorch with data preprocessing techniques: resizing, normalization, and one-hot encoding of masks.
Technologies: Python, PyTorch, CNN, Attention Mechanism Github: https://github.com/PhucLe9426/DUCK-Net-with-attention 5/2025 - 7/2025 Attention Tracker
Developed an Attention-based Transformer model for detecting prompt injections in natural language inputs.
Utilized the deepset/prompt-injections dataset to evaluate model performance on prompt injection detection tasks.
Implemented in PyTorch with attention mechanism analysis, including attention map processing and scoring for detection.
Technologies: Python, PyTorch, Transformers, Attention Mechanisms, Hugging Face, scikit-learn Github: https://github.com/PhucLe9426/prompt_tracker_cdmmm 1/2025 - 3/2025 Traffic Sign Classification
Developed a deep learning system using PyTorch and ResNet50 for traffic sign classification and segmentation.
Achieved high accuracy on 5-class image classification with transfer learning. Built a Streamlit app for real-time prediction with confidence scores and top-3 results. Technologies: Python, PyTorch, ResNet50, DeepLabV3, Streamlit, OpenCV, PIL, Transfer Learning, Semantic Segmentation, Image Classification
Github: https://github.com/PhucLe9426/Traffic-Sign- 1/2025 - 4/2025 Brain Tumor Segmentation
Built a VGG16-based model for brain tumor classification from MRI scans, integrated with RapidMiner for analytics.
Applied transfer learning to distinguish HGG vs. LGG tumors with high accuracy and automated data preprocessing.
Built web app with real-time image processing, statistical insights, and medical data export. Technologies: Python, TensorFlow/Keras, VGG16, FastAPI, RapidMiner, OpenCV, Pandas, scikit-learn Github: https://github.com/PhucLe9426/Brain_Tumor_SEG SKILLS
Data Analyst Python(Numpy, pandas, scikit-learn)
Deep Learning Neural networks with TensorFlow and Keras Natural Language
Processing
Text preprocessing, sentiment analyst, prompt injection detection Database SQL, Hadoop, MongoDB
© topcv.vn