Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I am a Bayesian statistics and time series researcher with hands-on experience building hierarchical time-varying parameter models with shrinkage priors and implementing scalable variational inference methods. I have a strong background in MCMC, stochastic gradients, and simulation-based evaluation, with applied forecasting experience on macroeconomic datasets. I am currently pursuing a Master of Statistics at ANU and have collaborated on research projects ranging from TVP estimation to priors and variable selection. I enjoy translating complex models into practical forecasting tools and look forward to advancing Bayesian methods for scalable inference in real-world data.…I am a Bayesian statistics and time series researcher with hands-on experience building hierarchical time-varying parameter models with shrinkage priors and implementing scalable variational inference methods. I have a strong background in MCMC, stochastic gradients, and simulation-based evaluation, with applied forecasting experience on macroeconomic datasets. I am currently pursuing a Master of Statistics at ANU and have collaborated on research projects ranging from TVP estimation to priors and variable selection. I enjoy translating complex models into practical forecasting tools and look forward to advancing Bayesian methods for scalable inference in real-world data.

YUEKUN MA

Data Scientist, Data Analyst, AI Engineer, +2





I am a Bayesian statistics and time series researcher with hands-on experience building hierarchical time-varying parameter models with shrinkage priors and implementing scalable variational inference methods. I have a strong background in MCMC, stochastic gradients, and simulation-based evaluation, with applied forecasting experience on macroeconomic datasets. I am currently pursuing a Master of Statistics at ANU and have collaborated on research projects ranging from TVP estimation to priors and variable selection. I enjoy translating complex models into practical forecasting tools and look forward to advancing Bayesian methods for scalable inference in real-world data.…I am a Bayesian statistics and time series researcher with hands-on experience building hierarchical time-varying parameter models with shrinkage priors and implementing scalable variational inference methods. I have a strong background in MCMC, stochastic gradients, and simulation-based evaluation, with applied forecasting experience on macroeconomic datasets. I am currently pursuing a Master of Statistics at ANU and have collaborated on research projects ranging from TVP estimation to priors and variable selection. I enjoy translating complex models into practical forecasting tools and look forward to advancing Bayesian methods for scalable inference in real-world data.

Available to hire

I am a Bayesian statistics and time series researcher with hands-on experience building hierarchical time-varying parameter models with shrinkage priors and implementing scalable variational inference methods. I have a strong background in MCMC, stochastic gradients, and simulation-based evaluation, with applied forecasting experience on macroeconomic datasets.

I am currently pursuing a Master of Statistics at ANU and have collaborated on research projects ranging from TVP estimation to priors and variable selection. I enjoy translating complex models into practical forecasting tools and look forward to advancing Bayesian methods for scalable inference in real-world data.

Skills

Python

Data Visualization

GitHub CoPilot

Experience Level

Python

Expert

Data Visualization

Expert

GitHub CoPilot

Intermediate

Language

English

Fluent

Work Experience

Researcher (Research Assistant) at Australian National University (ANU)

May 1, 2025 - December 1, 2025

Reviewed Bayesian Lasso and horseshoe shrinkage priors; compared computational implications (e.g., half-Cauchy hierarchy vs. conjugacy) and summarized MCMC strategies (hybrid Gibbs with MH vs. pure Gibbs via auxiliary-variable representations) to support stable inference in high dimensions. Investigated grouped and adaptive/group-structured horseshoe priors to incorporate structured sparsity for TVP models; formulated a grouped-shrinkage approach to improve interpretability under many coefficients. Implemented TVP estimation by extending the Bitto & Fröhwirth-Schnatter (2019) scheme to support horseshoe, grouped horseshoe, and adaptive grouped horseshoe priors in a unified workflow; developed a Bayesian variable-selection procedure that selects covariates over the full time horizon. Ran simulation studies to benchmark priors and selection behavior; applied the method to ECB euro-area inflation data (Feb 1994–Nov 2010, T=190) with 37 potentially time-varying coefficients.

Research Assistant at University of Melbourne (UniMelb)

December 1, 2025 - Present

Supported a project extending doubly stochastic variational inference (DSVI) from the Gaussian variational family to skew-symmetric distributions; formalized the ELBO objective and reparameterized latent-variable representations to enable scalable stochastic-gradient optimization with Monte Carlo estimators. Derived and documented gradient estimators for key variational parameters and investigated computational strategies for correlated posteriors, including precision-matrix parameterization and variance-reduced estimators to improve numerical stability in non-Gaussian settings.