Machine Learning Computer Science

Nguyen Hoang Hiep – Intern AI Engineer

***********@*****.*** 039*-***-***

Go Vap, Ho Chi Minh City, Vietnam

LinkedIn: linkedin.com/in/nhh04

ABOUT ME

A self-motivated and detail-oriented final-year Computer Science student with a strong interest in AI, Agents, NLP, and data-driven systems. Passionate about solving real-world problems through practical applications of machine learning and deep learning. Soft Skills: Critical thinking, teamwork, adaptability, fast learning, effective communication, problem- solving mindset.

EDUCATION & CERTIFICATES

University of Information Technology - VNUHCM (UIT) 2022 – 2025 Bachelor’s Degree in Computer Science – Graduated: July 2025 GPA: 3.5 TOEIC Listening & Reading: 915 / 990 December 2024 SKILLS & TECHNOLOGIES

Programming: Python, C++, JavaScript/TypeScript

Databases: MongoDB, PostgreSQL, Qdrant, MySQL

ML/DL Frameworks: PyTorch, TensorFlow, Scikit-learn, Huggingface, Transformers NLP Tools: underthesea, spaCy, pyvi, nltk, KeyBERT CV & OCR: YOLO (Ultralytics), OpenCV, PaddleOCR, EasyOCR Utilities: Langchain, GOOGLE-ADK, FastAPI, Pandas, Numpy, Matplotlib, Docker, A2A Version Control: Git, GitHub

PERSONAL PROJECTS

Helmet Detection and Rider Tracking using YOLO [GitHub] Built an object detection system to recognize motorcycle helmets in traffic images using YOLO-based deep learning models (via Ultralytics). Preprocessed and labeled the dataset, visualized detection results, and evaluated model performance using metrics such as mAP and IoU. Deployed the trained model using Streamlit for real-time web-based interaction and testing. UIT Admission Chatbot [GitHub]

Built a RAG-based QA system for student inquiries on majors and admissions. Combined query classification, semantic search (Qdrant), and LLM response generation. Used underthesea, KeyBERT, and Sentence Transformers for preprocessing and keyword–semantic fusion. Applied metadata filtering and prompt engineering to enhance relevance and response quality. UIT Data Science Challenge 2024 (Sarcasm Classify - 1st Runner-Up) [GitHub] Built a multimodal deep learning pipeline to classify sarcasm into four categories: TEXT-SARCASM, IMAGE-SARCASM, MULTI-SARCASM, and NON-SARCASM. Utilized Vietnamese NLP models

(BARTpho-v1, Vietnamese-SBERT, jina embeddings) and vision models (Swin Transformer, ViT Base). Integrated OCR (PaddleOCR, EasyOCR) and advanced fusion strategies. Problematic Internet Use Prediction – Kaggle Competition [GitHub] Participated in the Child Mind Institute challenge on Kaggle. Conducted data cleaning, feature en- gineering, and EDA using Pandas and Jupyter Notebook. Trained and evaluated models including LightGBM, XGBoost, CatBoost, and TabNet. Tuned hyperparameters via cross-validation and as- sessed models using ROC-AUC and F1-score.

1