Eric Stern
Email: eric.stern1@gmail.com Github: github.com/estern95 Phone: (303)-883-1810
Work Experience
Amazon, Data Scientist II - People Experiences & Technology Fall 2021 -Present Seattle, WA
Job Recommendation Systems
- Developed recommendation systems for associate roles within the Amazon job network to improve engagement of lost or at-risk candidates using content based and collaborative filtering approaches. The final model leveraged a neural network based collaborative filtering approach with constrained optimization.
- Orchestrated A/B tests to evaluate final design based on KPIs to account for candidate satisfaction and network fill ratio.
- Collaborated with software engineers and researcher scientists to create the final production version using AWS Sagemaker and Tensorflow.
- Job recommendation improved candidate satisfaction KPIs by 17% and improved fill in key regions by 11%.
Schedule Generation and Optimization
- Created optimization models to generate schedules for associates using mixed integer programming.
- Collaborated with software engineers to scale this to allocate 100,000 jobs per year
Labor Order Optimization
- Developed machine learning based control systems, using LightGBM and control theory to optimize the duration of job postings within the Amazon network.
- Built out the MLOps suite using Sagemaker pipelines and Cloudformation, enabling shadow testing, drift detection, and automated retraining.
- Designed and implemented A/B tests using Sagemaker experiments. The novel model improved labor order fill by 10%.
Guidehouse, Data Scientist - Advanced Solutions Spring 2017 - Fall 2021 Boulder, CO
Electric Vehicle Adoption Modeling
- Maintained an electric vehicle adoption and charger optimization model based on the Bass Diffusion model. The model has been used by 50 automotive and utility clients.
- Redesigned and deployed the model as a customer facing web application using a Plumber API and Shiny.
Energy and Demand Efficiency Evaluation
- Evaluated utility efficiency portfolios to ensure compliance with regulatory requirements.
- Performed econometric evaluations to estimate the effectiveness of behavioral efficiency measures such as smart thermostats, lightbulb rebates, and home energy reports.
Volt-Var Optimization Causal Analysis
- Designed experiments and analysis to estimate the effect of volt-var optimization programs on energy efficiency and energy demand using S-learners and econometric methodologies.
Education
University of Puget Sound 2013-2017 Tacoma, WA
B.S. in Economics
Honors Thesis: Modeling the Decoupling of Energy Consumption and Economic Growth
Relevant Coursework: Econometrics, Behavioral Economics, Experimental Economics, Math Econ, Linear Algebra.
Programming Languages & Tools
R (6 years), Python (5 years), Git (6 years), Shiny (3 years), Spark (3 years), SQL (6 years), AWS (4 years)