Corey Ducharme Email and Phone Number

Core software developer, data scientist, and machine learning engineer working on the digital twin simulation asset FutureScape. Full Stack Software EngineerCore developer on automating the use of the simulation tool FutureScape, converting it from a simulation engine to a fully-featured SaaS. Responsible for the multi-tenant database design (PostgreSQL), asynchronous back-end rest API and file services (Python, Node.JS, Java), custom user script management (Docker), cloud integration & DevOps (AWS) and front end (Angular, React). Data ScientistCore data scientist working on improving the capabilities of the simulation engine using recent advances in machine learning. Converted existing jupyter notebooks into high performance production ready python/sql scripts. Delivered quality analytics and interpretation in client facing roles.Machine Learning EngineerDesigned an end-to0end deep learning Variational Autoencoder (VAE) to produce synthetic population of agents being simulated. Novel design capable of handling both categorical and numerical attributes of the inputs in the same network. Implemented the design in TensorFlow 2 (python) as custom model class with custom layers.Technologies usedCloud: AWS : EC2, RDS, Batch, S3, ECR, Cognito, DynamoDB, Code pipeline, Athena, GlueProgramming Languages: Python, SQL, Javascript, TypeScript, Bash, R, JavaProgramming frameworks: FastAPI, Angular, Express, Node.js, Spring BootTools: Docker, gitDeep learning: Tensorflow

Performed survival analysis of industrial process to determine failure rates of machines using SAP Predictive Analytics.Programmed an R shiny data visualization of a product assembly line network optimization.

Developed a web based data visualization tool with Python (Bokeh) and JavaScript which allowed stakeholders to identify market expansion opportunities.Designed a statistical model that relates customers' demand and products' elasticities using hierarchical linear mixed effects which evaluated and guided potential market expansions decisions.Implemented and presented a distributed version control system (git) which simplified collaboration between researchers in the group.Set up an SQL database to combine and collect various sources of customer and logistic data for the research group.

Designed a novel clustering technique using spectral decomposition of time series to identify similar behavior in customer's demand.Programmed multiple forecasting models found in literature to compare to newly researched methods.Presented results to partner company.

Designed mechatronic characterization device of composites materials.Proposed a novel control method for the device to allow long distance and precise actuation.Final plans drawn in Catia V6.