732-***-**** Yonaton Heit Ph.D. Riverside, CA

ad2lis@r.postjobfree.com Data Scientist linkedin.com/in/yonaton-heit Technical Skills

• Knowledge in Programming: Python, C++, Perl, Fortran, Linux, Jupyter-Notebook, Anaconda, Git

• Data Management: SQL, Pandas, BiqQuery, Google Cloud Platform, Amazon Web Services, Sagemaker, Matlab, Excel

• Machine Learning: Scikit-learn, NLTK, TensorFlow, Keras, Deep Learning, XGBoost, ARIMA

• Miscellaneous Modeling Methods: Natural Language Processing, Demand Forecasting, Recommendation Systems, Time Series Analysis, Collaborative Filtering

Work Experience

• Senior Data Scientist June 2022—Present

XO/Vista Global Fort Lauderdale, Florida

– Developed and managed a dynamic pricing model for private jet flights, incorporating historical booking data and projected demand to optimize pricing strategies. The algorithm resulted in a projected revenue increase of approximately $63 million or 10% compared to the baseline in a 4 month period.

– Designed, built, and implemented a Gradient Boosting model to correct private jet prices from an external API. This model increased the predicted price’s R2 from 81% to 90%.

– Developed and deployed an AWS-based ensemble forecast pipeline to predict flight bookings, informing critical company functions such as aircraft availability and booking prices.

• Data Scientist Sept. 2019—March 2022

iHerb Irvine, California

– Designed, built, and implemented recommendation systems, programmed in Python and utilizing SQL queries, for our e-commerce website such as

* a statistical model to predict the time of repurchase in order to optimize customer repurchase recommenda- tions. A/B testing found a 17% lift in items added directly to cart compared to older versions.

* a trending keywords algorithm, which uses natural language processing to produce a list of keywords related to products that are treading based on a higher search frequency than normal variation. This algorithm worked for multiple languages and was used to introduce customers to new products.

– Developed and maintained demand forecasting model that utilized ARIMA to capture the trends in time-series data and Gradient Boosting machine learning to capture the spikes from promotions and seasonality for 30 thousand products.

– Developed a multiple language key-phrase tag algorithm for product reviews. This method utilized word embed- ding and clustering in order to pick out the top set of key phrases that appear in product reviews.

• Postdoctoral Researcher/Computational Chemist Sept. 2016—Sept. 2018 University at Buffalo Buffalo, New York

– Developed and implemented algorithms, written in Python, Perl, and C++, to model absorption and the MCD spectrum from vibrational modes and multi-configurational active spaces resulting in 3 publications.

• Graduate Researcher and Teaching Assistant Sept. 2010—March 2016 University of California Riverside Riverside, California

– Developed an original algorithm, written in C++ and Fortran, that determines and utilizes symmetry to shorten the computational time by 2-36 factors for the in-group code used by the team.

– Wrote algorithms to determine crystal expansion using statistical mechanics resulting in multiple publications and research, which contributed to a resulting $460,463 NSF grant to the group. Education

• University of California Riverside, Ph.D. in Computational Chemistry Sept. 2010 —March 2016 Riverside, CA

• Fairleigh Dickinson University, B.Sc. in Chemistry Sept. 2006 —May 2010 Madison, NJ

Honors: Magna Cum Laude