Data Scientist

Wen Fong Goh Email : *****@*******.***

http://goh.physics.ucdavis.edu Mobile : 530-***-****

http://github.com/wfgoh Location: San Jose, CA

Education

PhD Physics, University of California, Davis Sept 2018 MS (Research) Physics, Universiti Sains Malaysia Aug 2013 BS (Hons) Pure Physics, Universiti Sains Malaysia Aug 2010 Experience

Data Scientist Intern Aug 2018 - Present

EDF Innovation Lab San Jose, CA

Weather Model Forecast

Enhanced weather model forecast accuracy by developing two methods including statistical analysis and machine learning (Seasonal ARIMA, Linear Regression, Ridge, etc.) using Python.

Increased accuracy to help trading team enhance risk management. Accuracy > 90% using statistical analysis, MAError < 2% using machine learning.

Built a software tool in Python for business and stakeholders to use in real-time forecasting and trading. Electricity Market Price Forecast

Applied machine learning approaches (Random Forest, Gradient Boosting, Deep Learning, etc) to forecast electricity market price using millions of time-series data.

Created data visualization for forecasting using Python and Graph theory (pandas, matplotlib, networkx, etc).

Summarized and presented the results to business and stakeholders.

Engineering Intern May 2018 - Aug 2018

Western Digital Corporation San Jose, CA

Physics/Engineering Algorithms Development

Developed algorithms and code components in Python from Physics research papers that help accelerate spin Hall conductivity for complex materials.

Applied ab-initio calculations on AWS cloud computing to predict the spin Hall conductivity in alloys and multilayers metals for non-volatile memory SOT-STT MRAM applications.

Optimized the doping concentration and enhanced the current best spin Hall conductivity by 20%.

Graduate Student Researcher Jan 2015 - May 2018

University of California Davis Davis, CA

Computational Materials Research

Predict and understand the physical properties of complex crystalline materials using numerical rst-principle calculations on high-performance computing clusters.

Completed 2 research projects, majorly contributed to 4 NSF/DOE funding, published 5 rst-authored papers in peer-reviewed journals and presented at 9 conferences and scienti c meetings.

Helped experimental group on a NMR electric eld gradient measurements study of stressed superconductor by calculating the charge distribution using DFT. Results in a journal publication.

Supervised a student on a research project results in a journal publication. Skills

Python, Fortran, SQL, Pandas, Mathematica, Excel, Machine Learning, Scikit-Learn, TensorFlow, ARIMA, Time-Series Data Analysis, Predictive Modeling, A/B Testing, Data Mining, Linear & Logistic Regression, Bayesian Model, Random Forest, Gradient Boosting, Decision Trees, Deep Learning, Statistics, MapReduce, Hadoop, Spark, OpenCV, Linux, Distributed Computing. Publications

8 publications in peer-reviewed journals, including 4 papers in Physical Review B and 7 rst-author papers. 10+ presentations in conferences and scienti c meetings. https://scholar.google.com/citations?user=fxd4KJYAAAAJ&hl=en Awards and Honors

Fellowship, Universiti Sains Malaysia (Jan. 2011 - Dec. 2012)

University Half Colours (Fencing), Universiti Sains Malaysia (2009/2010).

University Half Colours (Athletics/Track and Field), Universiti Sains Malaysia (2009/2010).

Deans Lists, Universiti Sains Malaysia (2007 - 2010).

ICAM funded travel grant, Rice University (Nov. 2016). Other Machine Learning Projects

Predict Energy Gap and Stability of Novel Transparent Semiconductors

Performed exploratory data analysis on 2400 materials data sets to identify the distribution and features that correlate with the predictions.

Generated 80 important descriptors using my knowledge of expertise.

Picked the machine learning algorithm that gives the lowest RMS error on test set while considering the bias-variance trade-o .

Obtained a R2 score (accuracy) of 95% for the prediction of energy gap using random forest regression.

Boosted the accuracy to 89% for the prediction of stability (formation energy) using gradient boosted trees due to higher bias by random forest.

Movie Recommender System

Built a movie recommendation system using collaborative ltering based on users ratings.

Justi ed that matrix factorization is better than k-nearest neighbor method due to the nature of user-item interactions.

Web Scraping

Scraped job advertisements from Indeed, which requires interactions with HTML and Javascript using Selenium.

Spam Email Detection

Classi ed spam/non-spam emails using FastText neural network and support vector machine learning from public dataset.

*These projects are available on http://www.github.com/wfgoh and http://goh.physics.ucdavis.edu/datascience. LIST OF PUBLICATIONS

1. W. F. Goh and W. E. Pickett, Phys. Rev. B 98 125147 (2018).

\Coemergence of Dirac and multi-Weyl topological excitations in pnictide antiperovskites." 2. W. F. Goh and W. E. Pickett, Phys. Rev. B 97 035202 (2018).

\Survey of the Class of Isovalent Antiperovskite Alkaline Earth-Pnictide Compounds." 3. T. Kissikov, R. Sarkar, M. Lawson, B. T. Bush, E. I. Timmons, M. A. Tanatar, R. Prozorov, S. L. Bud’ko, P. C. Can eld, R. M. Fernandes, W. F. Goh, W. E. Pickett and N. J. Curro, Phys. Rev. B 96 241108(R)

(2017).

\Local Nematic Susceptibility in Stressed BaFe2As2 from NMR Electric Field Gradient Measurements." 4. W. F. Goh and W. E. Pickett, Phys. Rev. B 95 205124 (2017).

\Competing Magnetic Instabilities in the Weak Itinerant Antiferromagnetic TiAu." 5. W. F. Goh and W. E. Pickett, Europhys. Lett. 116 27004 (2016).

\A Mechanism for Weak Itinerant Antiferromagnetism: Mirrored van Hove Singularities." 6. W. F. Goh, S. A. Khan and T. L. Yoon, Modelling Simul. Mater. Sci. Eng. 21 045001 (2013).

\A Molecular Dynamics Study of Thermodynamic Properties of Barium Zirconate." 7. W. F. Goh, T. L. Yoon and S. A. Khan, Comput. Mater. Sci. 60 123 (2012).

\Molecular Dynamics Simulation of Thermodynamic and Thermal Transport Properties of Strontium Titanate with Improved Potential Parameters."

8. W. F. Goh and W. E. Pickett, in preparation (2018).

\Potential Topological and Thermoelectric Materials in Double Antiperovskites."

Contact this candidate