Developed and owned the core Instrument Control Software, interfacing between FW and the client UI. Built in Python/C++ with a single-threaded asynchronous architecture.Designed, owned, and led a team to develop a company-wide data visualization service, emphasizing fast data retrieval and rapid analysis, with a focus on demystifying our data.Created Data Engineering infrastructure with a Python ETL service, FastAPI API, and a neo4j database to enable future Deep Learning projects on high-density genetic data.Developed recipe engine and trained engineers to create and use recipes coordinating the operations of our benchtop equipment. This was custom compiled scripting language that directly integrated with our control software.Main technical interface/SME between FW engineering, Mechanical Engineers, Bioinformatics, and the Systems Integration to the instrument software team.

Designed and led a team to build a complete data pipeline, Postgres/Elastic data lake, andcustom Golang APIs to enable deep learning on customer data from autonomous retailenvironments. Revamped the original design to minimize monthly operation costs.Led CNN-based computer vision project to OCR nutrition label information from productappearance onboarding images.Developed in the Python based store automation backend in both new algorithm implementationand refactoring backend design. Improved new product onboarding time by 20x.Implemented CI/CD pipelines for multiple products, setting up automated image builds andtest runs through Github Actions and Jenkins.Set up new BI division, complete with custom D3, Python, and Tableau secured backend.

Designed Kalman Filter based real-time signal processing algorithms for maneuvering trackclass problems, implementing in server side applications. Designed and oversaw manufacture of UAS drones, implemented with unique payloads forpublic/private sector customers. Achieved 70% reduction in variance of key variables through novel implementations ofcutting edge signal processing research. Worked with customers to define and manage system-level requirements and successfullyconstrained the system to work above expectations and within requirements.

Developed high frequnecy PI control loop in Simulink for microsecond-scale tuning of optics, optimizing for a theoretical model-based maximum. This contol loop replaced an error-prone dither loop with the introduction of new optics metrology.Developed KPIs to identify high-speed error signatures in low frequency data, including a novel implementation of total variation limited regularised differentiation. Applied these KPIs against months of data to identify new failure signatures.Applied modern machine learning methods in multiple projects data exploration phases. Used Random Forest Regression and Neural Networks to develop much more powerful predictors, identifying otherwise unknown inflection points in long-term trends and providing guidance during development of human-understandable white box models.

I am responsible for the development of system level experiment design, execution, and analysis that focus on the increased power output of next generation EUV sources. I also develop and construct both software and hardware test tools focusing on generating diagnostic information on R&D machines. I have traveled as a expert technical adviser to aid local teams at customer fabs in the operation of next-generation EUV sources.

I assisted two graduate researchers in their research on the use of thermal plasma interaction with materials to develop better plasma models and create new silicon substrates for solar cell development. I also lead a research effort on DBD-plasma deposition on paper substrates and powder deposition for chewing gum manufacturing. I worked on redesigning a plasma torch apparatus to allow for a inert flame sheath.