Machine Learning Computer Vision

T R A N N H U T Q U A N G

076******* ************@*****.*** github.com/QuangTocDo

E D U C A T I O N

Ton Duc Thang University

Bachelor of Computer Science - GPA : 3.3/4.0

Thesis: Abnormal Heartbeat Detection Under Data Scarcity Aug. 2021 - Sep. 2025

WORK E X P E R I E N C E

P R O J E C T S

Abnormal Heartbeat Detection Under Data Scarcity Python, Pytorch Use MIT-BIH Arrhythmia dataset and preprocess ECG signals for multi-class classification

Combine PPSN and CNN to extract shapelet-based features and global patterns Achieve over 98% accuracy on test set with strong generalization across heartbeat types

FINPROS Investment Joint Stock Company (FINPROS JSC) Machine Learning Engineer Intern

Stock Prediction Python, Tensorflow, Matplotlib

Preprocess stock data from FPT, MSN, PNJ, and VIC, including scaling and splitting into train/test sets.

Train ML models (Linear Regression, XGB) and DL models (LSTM) for stock price prediction.

Evaluate and compare models using metrics RMSE, MSE and visualize predicted vs. actual prices for insights.

Sep. 2024 - Feb. 2025

RFM using Pyspark Python, Pyspark

Load customer transaction data into a PySpark DataFrame and calculate Recency, Frequency, and Monetary values using group operations. Use window functions to rank customers based on R,F,M scores, dividing them into groups (e.g., quartiles).

Combine the RFM scores to categorize customer segments and analyze their behavior efficiently at scale.

T E C H N I C A L S K I L L S

AI/ML & Computer Vision: Deep Learning, Transfer Learning, Image Classification, Object Detection (YOLO), Image Segmentation (U-Net, SAM2), Time Series Analysis Programming Languages: Python, Java, R, C/C++

Frameworks & Libraries: PyTorch, TensorFlow, Pyspark, scikit-learn, OpenCV, NumPy, Pandas, Matplotlib, Flask, FastAPI

MLOps & Deployment: Docker, ONNX, FastAPI, AWS

Developer Tools: VS Code, PyCharm, Jupyter Notebook, Git Annotation Tools: CVAT, Roboflow, LabelMe

UNet for Breast Cancer Segmentation Python, Tensorflow, Matplotlib, Flask Collect and pre-process aligned pair mask images and apply data augmentation in the Mias dataset

Implement and train a tailored UNet(backbone Resnet24) model with validation. Achieve an accuracy of over 90%

Deploy model on android app with Flask backend

Semantic Traffic Navigation in District 10 Python, Tensorflow, Selenium, Flask Use selenium to drawl traffic images on roads in district 10 Use Labelme to label data for segmentation task

Implement and pretrain Unet(backbone Resnet24) model Achieve an IOU over 80%

Deploy model on android app with Flask backend

Face Query with LSH on Android and Django Backend App Using the Facescub dataset with over 100,000 images of over 530 people Research the use of CNN models for extracting features from facial images. ImpImplement and pretrain a CNN (MobileNetV2) model with validation. Use cnn to extract features to input into lsh model in pyspark Deploy model on android app with Flask backend

Object Detection & Segmentation System YOLO + SAM2, ONNX, FastAPI Developed an end-to-end pipeline combining YOLO for object detection and SAM2 for image segmentation to handle complex visual tasks. Optimized and converted models to ONNX format for accelerated inference on both GPU and CPU.

Deployed the solution via FastAPI, delivering real-time responses under 100ms. Integrated RESTful API for seamless connection with external monitoring systems and applications.

Contact this candidate