Jiamin (Jasmine) He Email & Phone Number

Madison, Wi, Us

Pleasanton, California, Us

Pleasanton, California, Us

Pleasanton, California, Us

Boulder, Colorado, Us

Data science lead for owning end-to-end machine learning pipelines to build out the next generation of our eCommerce products with more than 1.5B events per day• Led a project team to develop a big data-driven product recommender system with product and user segmentation by using NLP and computer vision (the first recommender in the company)• Developed the end-to-end eCommerce product categorization pipeline (Test accuracy > 92%) by using BERT in TensorFlow and Docker to improve user experience• Led a project team for building an AI-powered Chatbot to boost customer engagement by being available 24/7 and reduce customer service costs• Created a product metrics dashboard with SQL in Looker to measure how customers adopt and use features across all the products in the company

London, Gb

AI Products & SolutionsPlayed a critical role in developing the OCR (combination of computer vision and NLP) and learner feedback capabilities for Pearson’s Aida Calculus app.• Built the end-to-end production pipeline for the OCR (TensorFlow based) model (CNN+RNN) using Google Cloud Platform.• Improved the OCR model performance by 66% through tuning hyper-parameters, implementing attention mechanism, and applying beam search algorithm.• Created the evaluation pipeline in Python to diagnose the machine learning model with Docker and Kubeflow.• Released the math handwriting recognition dataset on Kaggle to provide the data science community with resources for undertaking the challenging OCR task.• Adhered to the principles of scrum and an agile framework in a fast-paced, development environment (start-up environment).• Collaborated with front-end developers, engineers, data scientists, and product managers to communicate and align priorities with the leadership.

Carmel, Indiana, Us

Valuation & Projection, Life Valuation Department• Implemented Principle Based Reserving (PBR) (First team to implement PBR within the company) - Performed PBR trainings to peers voluntarily - Completed PBR actuarial report from scratch• Built valuation model for the completely new product and enhanced its efficiency in PolySystems• Improved model accuracy by fixing assumption coding errors in PolySystems• Performed Cash Flow Testing and presented the waterfall analysis results to management team

Carmel, Indiana, Us

Pricing & Experience Study, Health Product Management Department• Discovered a loophole in health products that opened the wellness benefits to fraudulent claims by using A/E method• Improved pricing assumptions for the group health products by using statistical analysis to identify patterns in policyholder behavior• Enhanced rate quoting engine’s accuracy by creating validation models• Automated regulatory filing’s projections and loss ratio calculations

Master'S Degree, Actuarial Science

Data Science Immerse Program

Bachelor Of Science (B.Sc.), Statistics