● Large scale, high-volume malware analysis platform

• Led a 6-month effort to update the codebase from Python 2 to Python 3 and improved test coverage and coding standards along the way

• Developed and enhanced the product by writing and debugging Python and JavaScript code for malware analysis tasks, web backend code, and web frontend code; increased detection rates of malicious patterns, performed analyses quicker, and distilled verbose results into actionable data

• Coordinated weekly deployments to production; ensured recently-added features functioned as expected and no issues appeared in production after deployment

• Created Python unit tests and Selenium-based UI tests to ensure code correctness and help prevent future bugs

• Constructively participated in code reviews and used Agile methodologies for working with eight other engineers to improve and maintain the product

● Adversarial machine learning testbed

• Leveraged state-of-the-art academic research on machine learning security to implement evasion and model-stealing attack algorithms in Python to be used against deep neural network models

• Streamlined the process for integrating new attack algorithms by using Docker containers to compartmentalize services and designed a Protobuf-based protocol for standardizing communication between master and worker containers

• Used React to build a web-based user interface that communicates with a Python-based web backend server over a RESTful API.

• Trained and tested machine learning models using TensorFlow 2.0 and Keras, to validate attack algorithms.

• Designed and implemented an abstraction layer in Python over the SciDB database system for storing and retrieving multidimensional datasets.