filter work by:    +All    +Data Science    +Design    +Sound    +Photo
Matthew Bejtlich © 2020
MATTHEW BEJTLICH

Matthew (b. 1991) is a data scientist, designer, researcher, musician, and photographer who is fascinated with how we can sustain relationships with each other and the natural world. He aims to foster communities, systems, strategies, and platforms that are more playful, honest, spontaneous, and nurturing at their core. His research exists at the interface of data science, experience design, human-computer interaction, participatory design, data ethics and data provenance, and creativity support tools.  
       For the past ten years, Matthew has been deeply invested in the global electronic music scene, producing electronic music as part of the audio-visual duo Dayspired. He is a current partner at Rhythm Section International (London, UK), where he is being mentored by record label founder Bradley Zero (DJ at BBC Radio 1, NTS Radio, Boiler Room). Soon he will also be exploring more minimal and atmospheric directions in solo work released under his own name.
       Matthew recently graduated with his MFA from Rhode Island School of Design (RISD) after receiving his MS in Data Science from Brown University. He was a graduate researcher at Brown’s Human-Computer Interaction Lab and the head teaching assistant for the class Machine Learning and Design. He is currently open to full time work and freelancing opportunities.


Connect
Email, Resume, Music, Instagram, LinkedIn, GitHub

Client list
He has worked or collaborated with Google, Holition, Warren Alpert Medical School of Brown University, British Fashion Council, Brown University HCI Lab, and the Naval Undersea Warfare Center.

Education
Rhode Island School of Design
MFA Graphic Design, 2021

Brown University
MS Data Science, 2018

University of Massachusetts Dartmouth
BS Electrical Engineering, 2015



MATTHEW BEJTLICH

Matthew (b. 1991) is a data scientist, designer, researcher, musician, and photographer who is fascinated with how we can sustain relationships with each other and the natural world. He aims to foster communities, systems, strategies, and platforms that are more playful, honest, spontaneous, and nurturing at their core. His research exists at the interface of data science, experience design, human-computer interaction, participatory design, data ethics and data provenance, and creativity support tools.  
       For the past ten years, Matthew has been deeply invested in the global electronic music scene, producing electronic music as part of the audio-visual duo Dayspired. He is a current partner at Rhythm Section International (London, UK), where he is being mentored by record label founder Bradley Zero (DJ at BBC Radio 1, NTS Radio, Boiler Room). Soon he will also be exploring more minimal and atmospheric directions in solo work released under his own name.
       Matthew recently graduated with his MFA from Rhode Island School of Design (RISD) after receiving his MS in Data Science from Brown University. He was a graduate researcher at Brown’s Human-Computer Interaction Lab and the head teaching assistant for the class Machine Learning and Design. He is currently open to full time work and freelancing opportunities.


Connect
Email, Resume, Music, InstagramGitHub, LinkedIn 

Press/Features
Apple Music
British Fashion Council  
Bolting Bits
Harper's Bazaar
Indie Shuffle
Stereofox
RISD Grad Show   
WWD

Client list
He has worked or collaborated with: Google, Holition, Warren Alpert Medical School of Brown University, British Fashion Council, Brown University HCI Lab, and the Naval Undersea Warfare Center.


PAINTING CLASSIFICATION WITH CNNS

Within recent years, museums and online art collectives around the world have started to digitize their art collections. In order to help support educators, curators, and archivists to archive and find interesting correlations between artwork, I helped to design a deep learning model.
        A database was formed with over 15,000 web-scraped images, each having attributes like (e.g. date and artistic style). A total 12 total artistic styles and 28 artists were represented. A convolutional neural network (CNN) was trained to predict style and artist, with a validation classification accuracy of 40% for style and 53% for artists. A simple visual interface was made to allow a user to run new (untrained) images through the model.
        The research was conducted under the guidance and mentorship of Dr. Serre from the Cognitive, Linguistic, and Psychological Sciences  (CLPS) lab at Brown University. 

Client: Self
Date: 2017
Class: Data and Computational Science
Instructor: Dan Potter, PhD


Matthew Bejtlich © 2020