Michael Cho Email & Phone Number

San Francisco, California, Us

Working for the Delivery Matching team

San Jose, Ca, Us

• Created new machine learning features to add to our recommendation system and rank similar items on the view item page by utilizing Scala and Spark for offline ETL, data analysis, evaluation, and online inferencing.• Finetuned a Siamese Bert model in Pytorch to increase performance on title similarity scores for different languages.• Utilized the Springboot framework to incorporate and test APIs for feature extraction.• Transitioned data pipelines from Jenkins to Airflow.• Trained a graph neural network model on GCP to produce a new recall set based on item coviews.

Armonk, New York, Ny, Us

• Architected and productionized a customer attrition model using AWS services such as Sagemaker, Stepfunctions, and Lambda for a medical device company.• Developed overall data model and created SQL and Pandas scripts to read and write data to Snowflake.• Created Docker images to use in conjunction with Batch for data pipeline.• Utilized Gitlab to automate the CICD process for model code packaging and QA.• Defined IAM roles for services and users and created Cloudformation templates for reference architecture.

Armonk, New York, Ny, Us

• Managed the development of a data pipeline for a state health department’s COVID operations and analytics workstreams, handling AWS Glue workflows, rearchitecting AWS services, augmenting the testing suite, and incorporating new data features in partnership with stakeholders.• Utilized PySpark to perform ETL for customized client extracts and data products, as well as to profile terabyte sized Hive tables for an automotive manufacturing company.• Created a data pipeline in support of a customer record linkage algorithm in AWS, incorporating tools such as Glue, S3, Spark, SNS, Redshift and Tableau to automate serverless daily workflows, send job notifications, and feed dashboards for a regional car dealership.• Automated the weekly migration of data from client’s on-prem sql server to Redshift using AWS Data Migration Services.• Managed the end to end design of overall solution by defining AWS services, detailing sizing requirements and cost considerations, creating data model, overseeing model development, and moving pipeline from development to production.• Developed a customer propensity model, using aggregated historical data, to score potential vehicle buyers which resulted in a 30% sales hit rate.

Armonk, New York, Ny, Us

• Accessed Weather Company and Google Map APIs to incorporate weather and geolocation data in a Python application for a national retail chain.• Created data model, logic, and testing suite to populate CouchDB NoSQL databases for email notifications and UI backend.• Developed a production multi-objective optimization model using Cplex to improve delivery time, quality, and waste of product assembled by restaurant workers for a fast food company.• Incorporated IOT mqtt messaging protocol to capture and structure real time streaming POS data into SQL databases.• Used Docker for continuous delivery of code updates and run simulations.• Created a Flask REST api to simulate an accelerated clock for end to end testing.

Moline, Il, Us

• Created a customer segmentation scheme of a recent acquisition by k-means clustering and developed a Tableau dashboard for the Tactical Marketing team to pursue sales leads.• Identified factors that leads customers to purchase compact utility tractors by testing classification models in caret and sklearn.• Scraped webpage data on competitors’ product line using the Beautiful Soup library in Python.• Converted a SAS job onto the Hadoop platform using big data tools such as Sqoop, Hive, and Oozie, reducing job time of a marketing distribution list from 13 hours to a half hour.

Chicago, Il, Us

• Joined SQL data about customers and flights in addition to web click data from Adobe Analytics to build a Random Forest model, creating a predictive model for product purchase.

Coral Gables, Florida, Us

• Created a CART model that reduced the review rate of an online retailer’s sub-channel by 30% by identifying more efficient fraud rules.• Validated production data with clients for ETL processes according to data mapping documentation.• Created SQL reports and presentations detailing descriptive statics on client KPIs and fraud trends.

Master Of Science In Analytics, Data Science

Master Of Management Studies

Bachelor Of Arts, International Relations Major & Economics Minor

Education record

Education record

Education record