data analyst

Yawei Qin Email : ******@*****.***

Mobile: 510-***-****

EDUCATION

Ph.D. University of California, Riverside GPA 3.89/4.0 Riverside, CA, USA Department of Physics & Astronomy Sept 2017 – June 2023 B.S. University of Science and Technology of China GPA 3.53/4.3 Hefei, Anhui, China Department of Physics Sept 2013 – June 2017

TECHNICAL SKILLS

Languages/Dev Tools:Python, SQL, C, Linux, Matlab, Mathematica, Cuda, Git, Gitlab Libraries : Pandas, Numpy, Scipy, Pytorch, Tensorflow EXPERIENCE

Quantitative Analyst Intern Jun 2022 – Aug 2022

Data analysis and Algorithm Development O’Neil Global Advisors

• Conducted data analysis on large time series datasets using statistical techniques to identify patterns and relationships between different factors and security performance for the FactorFarm project

• Utilized AWS Cloud Databases and GitLab to efficiently manage and analyze large datasets

• Developed skills in statistical analysis, data visualization, and presentation Research Assistant Sept 2017 – Now

Quantitative Model and Algorithm Development UC Riverside

• Developed an algorithm for outlier detection in time series data

• Used Bayesian inference to learn signal distributions and wrote stochastic differential equations

• Developed a numerical module to solve the Fokker-Planck equation using sparse matrices

• Used parallel computing on HPC to speed up computation on large data sets PROJECTS

Article Classifier Python Linear Classification Source Code

• Applied Support Vector Machines (SVMs) to accurately predict the subject of an article, achieving a 62% accuracy rate

• Developed and implemented a bag-of-words approach to convert article abstracts and titles into high-dimensional vectors for multiclass classification using SVMs Credit Card Fraud Detection Python LightGBM Source Code

• Implemented a highly accurate classification model for credit card fraud detection using Light Gradient-Boosting Machine (LightGBM), achieving an impressive score of 0.8 out of 1

• Leveraged Optuna, a hyperparameter optimization framework, to efficiently identify the optimal parameters for the model and further improve its performance Natural Language Processing with Disaster Tweets Python Natural Language Processing Source Code

• Developed and implemented a highly accurate classification model for identifying tweets related to real disasters using the pre-trained model ELECTRA as a text discriminator

• Fine-tuned the model with tweet data, leveraging techniques such as transfer learning and data augmentation to enhance its performance

• Utilized GPU acceleration to significantly reduce the training time of the model and achieved 100x performance improvement compared to CPU algorithms RELATED COURSES

• Fundamentals of Quantitative Modeling • Machine Learning Foundations: A Case Study Approach

Contact this candidate