Machine Learning Financial Data

Vu Duc Thanh

+84-096*-***-*** ***********@*****.*** https://www.linkedin.com/in/vuducthanh2611/ https://github.com/vuthanh2611 WORKING EXPERIENCE

Data Analysis –FiinGroup, Vietnam (06/2023- now)

• Successfully designed and managed an advanced ETL system, efficiently transforming a vast database comprising over 2 million Vietnamese companies into refined features, providing valuable business insights

• Developed Machine Learning Solutions for Risk Management, Including Corporate Credit Score (Credit risk), Corporate Failure score. This model was built on an extensive database of FiinGroup and can Score ~ 2 million Vietnamese enterprises and can apply to various use cases of both Corporate and Financial Institutions.

• Created data-driven reports in order to monitor Industry trends, business cycles, and to find top performers in the industry. Quantitative Consultant –WorldQuant, Thailand (12/2022 - 06/2023)

• Created quantitative model (alpha) using various data type including both financial data like revenue, profit and non- financial data such as sentiment data, news data, research funding data for trading equities in USA market

• Achieved an exceptional 'Gold' level status after successfully mastering ten alpha algorithms.

• Achieved a remarkable feat by securing the top 1% position outperforming over 31,000 entrants in the prestigious Worldquant competition

Data analysis –Habour Space, Thailand (08/2021- 12/2022)

• Utilized statistical methods and data visualization techniques to generate meaningful insights from financial data, aiding in informed decision-making for budget allocation and resource management.

• Leveraged machine learning insights to assist school counselors and advisors in providing tailored career guidance to students, aligning their interests with potential job opportunities.

• Utilized clustering and predictive modeling techniques to determine the type of students the school should target in their recruitment efforts to maximize student success and retention. Research Assistant – Ulsan National Institute of Science and Technology, South Korea (09/2019- 08/2021)

• Collected and compiled large datasets on concrete material properties, including material components, strength, and other relevant factors. Cleaned and preprocessed to ensure data accuracy and consistency for analysis.

• Developed and trained machine learning models to predict concrete strength based on the collected data with R2 around 98%

Project Management Engineering – Zamil Steel, Vietnam (03/2019-08/2019)

• Developed a road map for projects including mile stone, scope statement from initiation till close out with 80% accuracy

• Monitored and coordinated with other departments to deliver more than 10 projects to customer TECHNICAL SKILLS

Programming languages Python, R, Matlab, C++, Java, SQL, Casandra, MongoDB Libraries

Platforms

Numpy, Pandas, Scikit-learn, Tensorflow, Keras, pytorch, Matplotlib, Seaborn Flask, Scene builder, jupyter notebook, AWS sagemaker EDUCATION

M.S. Data Science, Habour Space @ University of Thai Chamber of Commerce, Thailand (GPA: 95/100) August 2022 M.S. Civil Engineering, Ulsan National Institute of Science and Technology, South Korea (GPA: 4.0/4.3) August 2021 PROJECTS

Emotion detection AWS SageMaker, tensorflow, python, numpy, keras, git

• Successfully designed and deployed an end-to-end machine learning pipeline on Amazon SageMaker in AWS,

• Achied over 80% accuracy in predicting student emotions based on the FER dataset. Poem generation based on the tale of Kieu LSTM, tensorflow, python, numpy, keras, git, flask, beautiful soup

• Utilized Beautiful Soup for text data collection.

• Created a Telegram chatbot powered by LSTM to generate unique poems based on user input like poem length and keywords.

• Github: https://github.com/vuthanh2611/poem_chatbot.git Object classification pytorch, vgg16, resnet18, python, git

• Created CIFAR-110 dataset by merging CIFAR-10 and CIFAR-100, implementing advanced augmentation for enhanced object classification

• Leveraged VGG16 and ResNet-18 architectures to perform precise object classification on the CIFAR-110 dataset

• Conducted meticulous experiments to fine-tune model hyperparameters and rigorously compared VGG16 and ResNet-18

• Employed transfer learning to adapt pre-trained VGG16 and ResNet-18 models to the CIFAR-10 dataset

Contact this candidate