Timothy Praditia

Data Scientist

Hi there! I am Timothy, a Data Scientist currently working for Bayer. My role involves supporting the digitalization of manufacturing processes and leading data analytics projects to deliver data-driven insights for internal stakeholders' strategic decision-making. Before that, I finished my PhD at the University of Stuttgart, where I did my research that focuses on the intersection between deep learning and physics, mainly to ensure the model's reliability as a tool to assist decision-making processes. In addition to machine learning and deep learning, my skills and experiences include (Python) programming, data visualization, cloud computing, machine learning, and statistics. I am also a huge fan of Juventus (the Italian football club, I even made a small analytics project on the team's performance with fellow Redditors).

Learn More

CV

My complete curriculum vitae.

Portfolio

Collection of my projects, both academic and for fun!

Publications

List of my publications and talks given in conferences.

Teaching

Courses that I teach and student projects that I supervise.

Projects

PDE Benchmark for Scientific Machine Learning

Generation of benchmark datasets based on various partial differential equations, and setup of deep learning models as baseline for the scientific machine learning community.

Finite Volume Neural Network

Replacing elements in a Finite Volume model with learnable neural network modules improves prediction accuracy, model interpretability, and generalization ability.

LSTM for Water Level Prediction

Development of an LSTM model to predict the swamp water level in an ecologically sensitive area, affected by nearby groundwater extraction and other hydrological processes, with additional physics-based constraints.

Thermochemical Energy Storage Modeling

Development of a surrogate model for a thermochemical energy storage system using NARX, equipped with physics-based regularizations.

Multiscale Modeling for Fractured Geothermal Reservoir

Simulating fractured geothermal reservoir models more efficiently by solving in coarse scale and reconstructing the solution into fine scale.

Juventus Analytics

A free-time collaborative project that collects and visualize data to analyze the performance of my favorite football club.

Festo Coding Challenge

A link to my code repository, that I wrote for the Festo Coding Challenge 2020.