*(NTU, GT) = (National Taiwan University, Georgia Institute of Technology)
Technical Skills
Programming Lang. / Tools : , R, / / + , , ,
Analytics Tools:
EDA and Data Preparation: Pandas, Polars, Dask, (Py)Spark
Visualization: Matplotlib, D3.js
Statistical Modeling and Machine Learning:
General-purpose Machine Learning Models: Sklearn
Deep Learning: Pytorch, HuggingFace
Functional / Tensor / Network / Textual / Time Series Data: Scikit-FDA, TensorLy, NetworkX, Statsmodels
Optimization: PuLP, CVXPY
Working Experience
Data Analytics Project Practicum Intern2024/09-2024/12PythonPowerBITime Series Analysis (@MedTranso Go + GT)
A practicum project sponsered by MedTrans Go, an end-to-end data analytics tasks, aiming to quantify and analyze tech updates on sales, as well as identifying other underlying patterns of requests.
Teaching Assistant and Research Assistant2019/08-2023/05 (@NTU, Math Department)
Hosted, instructed weekly homework, quiz review sessions & office hours for Calculus, Abstract Algebra as a TA
Collaborated with Prof. Wu-Yen Chuang to work on problems in algebraic geometry as an RA, organized study
groups, hosted weekly reading seminars on topics in algebraic geometry and algebraic topology
Selected Projects (@Georgia Tech)
This page provides an overview. A more detailed description is given in this page.
A project exploring ways to scale the traditional user-based collaborative filtering (UBCF) to larger datasets, via content-based (CBF) approaches - users subsampling and item clustering
A Python package published on TestPyPI, for simulation, calculation of statistics, visualizations of a particular theoretical flu-spread process
Miscellaneous Experience
3DGauCIM: Accelerating 3D Gaussian Splatting via Digital Compute-in-Memory for Real-Time Edge Rendering2024/08-2024/11publication
Fourth author in a publication submitted to DAC (Digital Automation Conference)
Helped producing rigorous key 3D visuals illustrating main ideas using Matplotlib, and modularized it as a Python dataclass for the paper’s main authors to easily make tweaks on it (such as switching on and off the plot components, using other color schemes or sizes)
Simplified and streamlined the mathematical formulations of Gaussian splatting (both 4D and 3D temporal)