Tianyu Ma

linkedin.com/in/tianyu-ma-96ab84288

Northeastern University College of Professional Studies, Boston, MA, USA 

1. System event log monitoring and anomaly detection

• This learning model can automatically identify and distinguish normal, suspicious, and malicious behaviors by analyzing system event logs.
• Mainly used random forest model in Python.

2. Provided ML4 Cyber company with a high-precision model for identifying software vulnerabilities

• Comparison of the Multilayer Perceptron (MLP) Model and the Convolutional Neural Network (CNN) Model
• Focused on using SGD and Adam optimizers to optimize the model

3. Provided Patient Power company with a predictive model for treating TBI patients

• I performed complex raw data cleaning
• I studied some of the key variables that had the greatest impact on the results and compared the accuracy of five models, and finally decided to use the Gradient Boosting Model (GBM) prediction model.

Nanjing University of Traditional Chinese Medicine, Nanjing, Jiangsu, China 

Research on the Production Process of Trimetazidine Dihydrochloride Modified Release Tablets.

• Preparing appropriate mobile phase and observe the product chromatogram using HPLC instrument
• Using data analysis software to process and interpret data, such as peak identification, peak area integration, calibration, etc.