Matthew Simpson Email & Phone Number

Raleigh, Nc, Us

Raleigh, Nc, Us

Raleigh, North Carolina, Us

- Led development team in the planning, design, and implementation of the MVP, releasing the first version within 2 months and landing our first customer shortly thereafter. - Architected the system backend, designing NoSQL databases and RESTful API endpoints, setting up pubsub message brokering and asynchronous task queues between services, and offering real-time event notifications with WebSockets. - Improved horizontal scalability by using Docker to containerize deployments, spinning up additional images during periods of high load. - Decreased API response times by 82% by implementing a database caching layer with Redis for frequently queried data. - Created the first open-source Object Document Mapper (ODM) for Firestore. - Technologies used include React, Firebase, Express.js, PhoneGap/Cordova.

Madison, Wisconsin, Us

- Improved performance of system file watcher by over 400% using multiprocessing to distribute and process incoming scan results in real-time across multiple workers.- Doubled file transfer throughput by orchestrating multithreaded job queues and chunked uploads to reduce network idling in IO-bound tasks.- Reduced client load by 30x using a SQLite cache to index and document filesystem layouts and content, thereby reducing disk reads and file lock contention.- Achieved 99% uptime through a complete re-write of the Windows service in Python & C. Application previously was hard crashing 3-4 times/week on high-priority customers.

Raleigh, North Carolina, Us

The Magneto-Optic Kerr Effect (MOKE) occurs when the left circularly and right circularly polarized components of linearly polarized light are projected onto a sample and reflect with different amplitudes and rotations from each other. This occurs in birefringent samples due to the different indices of refraction for the left and right circular components. The use of MOKE to characterize the surface magnetization of a sample was proposed in order to create a contactless scheme for magnetic field effect (MFE) spectroscopy and has important implications in Spin Caloritronics. We construct a loopless Sagnac Interferometer to demonstrate the ability of this approach to measure MOKE across an array of magnetic fields for Bi-YIG, NiFe, and LSMO films. This setup can be used for 2D microscopy and projects involving the Spin Seebeck Effect.

Colorado Springs, Co, Us

• Designed machine learning models with Google Tensorflow, libSVM, and MATLAB to replace automatic algorithm selection modules in scientific software.• Engineered meta-algorithms based off neural networks that could select optimal algorithms with 97% accuracy, leading to runtime improvements by margins of up to 75%.• Constructed mathematical functions that, in conjunction with the aforementioned AI models, decreased average computational error by 27%.• Published and presented initial research findings at the IEEE International Conference on Machine Learning and Applications. Further developments have led to even greater metrics than those mentioned in the paper (attached below).

Auburn, Al, Us

Results from this REU were published in the International Journal of Mathematics and Computer Science and the Alabama Journal of Mathematics.

Morrisville, Nc, Us

• Increased throughput of CAD software by 30% by creating a .NET application to monitor user processes and respond to events by performing realtime “tweaks” to memory paging, thread management, registry values, and software settings.• Achieved 26% higher benchmark scores for Solidworks and AutoCAD products as compared to results from Tom’s Hardware.• Compiled results into a database for use in customer facing products.

Bachelor Of Science - Bs, Computer Science (Bs), Physics (Ba), Mathematics (Bs)

Education record