Matthew Phillips Email & Phone Number

Minneapolis, Mn, Us

Boston, Massachusetts, Us

I lead the Deep Learning Research team at Elucid. We develop machine learning algorithms to interpret and understand processed computed tomography (CT) images of vasculature, coronary and carotid arteries in particular. We create, train, and deploy these models on our multi-GPU cluster. Our work in classifying vessel cross sections by pathology was presented at SCCT 2021, and was critical to the success of our Series B funding round. We also are aggressively researching solutions in precision medicine and the 'omics space, where I have contributed to 7 issued patents.

Boston, Massachusetts, Us

I build systems to solve problems in Computer Vision, NLP, and Signal Processing, as well as models of complex data sets in healthcare and other domains. I use modern Deep Learning technologies and AWS cloud services to accomplish this, along with traditional 'shallow' ML methods, optimization, and highly performant C++ code.Past projects:- Developed mortality and length-of-stay prediction models for the healthcare industry, and used O/E ratios to enable statistically valid cross-center comparisons.- Developed a system to schedule eating events for elite athletes for Brainbuild (www.brainbuild.io), using Reinforcement Learning.- Developed an optimized system for accurate measurement of fluid flow for Kesem Health (www.iuflow.com).- Developed a website and server backend for customer data storage and web-based analysis.

Indianapolis, In, Us

I created machine learning algorithms which address core problems in health care and wellness. One key project included using document denoising and text-based document classification for better understanding of electronic healthcare records (EHRs) (see link below). Additional projects included the use genomic data to determine tumor site-of-origin, and biomedical image segmentation. I dealt with numerous storage formats such as DICOM, FHIR, HL7, and ICD9/10 encodings along the way. Most recently we did causal inference analysis on the Precision Cancer Cohort A data we received from the VA as part of their AI-able Data TOP Health sprint, and presented these results at the SAMSI Causal Inference Opening Workshop in Dec. 2019.

Clifton Park, New York, Us

I developed Machine Learning and Computer Vision algorithms. Projects included the AFRL VIGILANT project, for which I used deep neural networks to perform supervised and unsupervised classification of satellite imagery (presented at the MSS Joint Conference Oct. 2018); a NASA SBIR grant on land cover classification where I was Kitware's technical lead (hyperspectral image segmentation); and a project analyzing and classifying decades of recently-digitized industry data (denoising algorithms, object detection). I also used aerial drone imagery to count and classify orchard trees. The paper from my satellite imagery project as well as another project using GANs for few-shot learning I contributed to were both selected as top 10 papers in the MSS conference and were presented at a followup joint conference in Feb. 2019. Outside of work, I organized multiple tech meetup/groups while at Kitware, serving as co-organizer of the Deep Learning RTP group (https://www.meetup.com/Deep-Learning-RTP/) from 2017-2018. In late 2017 I founded the Triangle Computer Vision and Image Processing Group, as well (https://www.meetup.com/Triangle-Computer-Vision-and-Image-Processing-Group/).In Feb. 2018 I delivered a talk on Generative Adversarial Networks to our local IEEE affiliate, and I was also a speaker at Duke's inaugural Triangle Machine Learning Day Event on April 3, 2018.https://www2.cs.duke.edu/induke/tmld/2018/index.php

Tempe, Arizona, Us

I developed software which ran the manufacturing process for Invisalign aligners. This was mostly in C++ but we also used some Python for testing and data visualization. Work included the creation and implementation of algorithms from computational geometry, optimization, and system design and architecture, as well as unit and functional testing. On the organizational side, I founded and ran the weekly Math/C++/Machine Learning reading group at Align. While at Align, I also founded and organized the Triangle C++ Developers Group, generously sponsored by Align.https://www.meetup.com/Triangle-C-Developers-Group/I used to hang around Stack Overflow and answer C++, algorithms, Qt etc. questions, but I haven't been active there in several years (http://stackoverflow.com/users/512947/matt-phillips).

Developing software applications in C++ for the biomedical industry. I was a senior developer on a large-scale project in the medical devices space, where developed and implemented algorithms for user input recognition, began construction of the GUI, and integrated these with the middleware.

Computational modeling in the lab of Stephen Lisberger, Department of Neurobiology. We are studying information processing in the cerebellum, specifically the significance of synchrony in Purkinje-Cell simple spikes. Initial results from this work have been presented at the Society for Neuroscience meeting, 2014.

New York, Ny, Us

Single-unit primate electrophysiology in the lab of Michael Goldberg, Department of Neuroscience. We studied the relationship between eye movements and the representation of visual space in primate parietal cortex. We also addressed these questions with monkey and human psychophysics, in the latter case including normal and patient populations. This research led to numerous presentations and one publication so far.

New York, Us

Oculomotor and visual psychophysics, specializing in visual search, in the lab of Jay Edelman. We monitored eye movements with a video eye tracker.

Post-Doctoral Fellowship, Neuroscience

Bachelor'S Degree, Philosophy/Mathematics Double-Major

Machine Learning

Doctor Of Philosophy (Ph.D.), Philosophy, Cognitive Science