Kevin's headshot

Hi, I'm Kevin McVey

I'm a computer vision specialist and engineering leader with experience bringing camera-centric products to market. Today I'm at Nautilus. Previously, Glowforge.

I'm also the Education Lead at Bandit Theater where I've taught the joyful art of improv comedy to hundreds of students.

My work has been on Mars, TEDx, and CBS's The Price is Right. One time I got trapped in an elevator and it became news.

Thanks for visiting my portfolio, I'd love to meet you!

Email · Resume · LinkedIn

Work Experience

Nautilus Biotechnology

Manager, Image Processing · Apr. 2024 - Present
Staff Software Engineer & Tech Lead, Image Processing · Oct. 2022 - Apr. 2024
  • Player-coach leading Nautilus's Image Processing Team; responsible for single-molecule detection on Nautilus's next-generation proteomics platform with a throughput of billions of events per hour.
  • Architected Nautilus's terrabyte-scale data collection, processing, and warehousing pipeline, led and project managed a cross-functional team of 10 software engineers, managers, and bioinformaticians in implementation across both instrument and cloud.
  • Led the development of on-instrument software, hardware test fixtures, and cloud-based data warehousing for optics and reagent quality control.
  • Developed distortion modeling and image registration techniques to enable assays with hundreds of cycles.

Glowforge

Senior Software Engineer & Tech Lead, Research Engineering · May 2018 - Oct. 2022
Software Engineer · Feb. 2016 - May 2018
Associate Software Engineer · Sep. 2015 - Feb. 2015
  • Led interdisciplinary teams (4-5 engineers) in prototyping and productization of computer vision features for Glowforge's cloud-controlled CNC laser cutter & engraver.
  • Awarded multiple US patents for new inventions in the creative tools / fabrication space.
  • Designed, trained, and evaluated deep learning models in the spaces of semantic segmentation, object detection, and classification.
  • Built training datasets of over 1MM images using crowdsourcing and automated annotation with sensors.
  • Developed algorithms for many customer-facing computer vision features. Areas include 3D reconstruction, multi-view geometry, feature detection & matching, camera calibration, and localization.
  • Scaled computer vision services on Google Cloud to support over 100,000 requests per day.
  • Designed Glowforge's CNC-optimized vector graphics format and built a corresponding web service that converts millions of customer-designed SVGs and PDFs per month.
  • Developed robotic factory calibration tools for manufacturing precision camera hardware at scale.
  • Mentored junior engineers, wrote technical documentation, and made frequent company-wide presentations.

Amazon

Software Engineer, Appstore Developer Experience

July 2014 - Sept. 2015
  • Architect of multi-tenant data collection and reporting pipeline used to track worldwide Appstore usage.
  • Sole developer of public-facing Appstore API that grew Appstore catalog by 20,000 apps.
  • Mentored an intern.

Software Development Engineer Intern, AWS Infrastructure

May 2013 - Aug. 2013
  • Built a Ruby On Rails application with Apache Solr to track the worldwide distribution of over 100,000 AWS assets.
  • Winner of 2013 Intern Coding Competition

BAE Systems

Virginia Microelectronics Consortium Research Intern

May 2012 - Aug. 2012
Education

University of Virginia

B.S. with Distinction: Computer Engineering & Minor: Architecture

2010 - 2014

Thesis: Personalization in SCOT: How User Groups Redefine Closure and Alter Product Lifetimes

Skills

Languages

Python, C++, JavaScript, Ruby, Java, C, Assembly, Rust, Bash, HTML

Tools

OpenCV, Tensorflow, Ceres-Solver, Numpy, Pandas, CUDA, Git, NodeJS, Flask, Rails

Cloud

Kubernetes, Docker, Google Cloud Platform, MySQL, Redis, AWS

Patents

Systems and Methods for Machine Vision

USPTO Application 18/670,575

Filed 2024-05-21

Methods for producing and utilizing a geometric camera calibraton created using randomly located fluorescent objects in a patterned flow cell.

Edge Detection for Computer Numerically Controlled Fabrication

USPTO Application 17/668,988

Filed 2022-02-02

Semantic and instance segmentation of materials in a CNC machine using one or more cameras. Segmentation is used to automate many advanced fabrication techniques including: object packing, storage and recall of scrap materials, adaptive tool pathing, and detecting the isometries of moving materials.

Multipoint Distortion Correction

USPTO Provisional Application 63/239,460

Filed 2022-01-14

Fast estimation of 3D depth maps of warped CNC materials using data collected by one or more cameras and/or sparse lidar. Maps are used to render high-accuracy 2D projections that rectify warps in the material in order to match the coordinate system of a moving CNC tool.

Computer Numerically Controlled Fabrication Using Projected Information

USPTO Application 17/133908

Filed 2020-12-24 ★ Granted 2023-08-29

Projecting structured light in various configurations onto the subject material in a CNC machine. Images of the structured light are used to infer various characteristics of the material including but not limited to size, shape, thickness, warps, encoded data, etc. Data may be sparse or dense in different implementations.

Fabrication with Image Tracing

USPTO 11,249,456 B2

Filed 2017-11-27, ★ Granted 2022-02-15

Automatic conversion of hand-drawn images into CNC tool paths. One or more cameras and/or lidar are used to locate drawings in 3D space and infer instructions.

Calibration of a Computer-Numerically Controlled Machine

USPTO 11,137,738 B2

Filed 2017-11-27, ★ Granted 2021-10-05

Multi-view calibration of (any combination of) overhead and moving cameras, lidar, laser optics, lighting, and moving tool head in a CNC machine. Method also produces tools to convert between each subassembly's coordinate systems. Self-supervised machine learning used to automate / bootstrap certain implementations.

Publications

The Electronic In-Patient Progress Note: Less Is More

L. Colligan, C. Coleman, S. James, L. Dobry, K. McVey, and S. Borowitz

American Medical Informatics Association (AMIA) Annual Symposium, 2013

This work presents an evaluation and redesign of Electronic Health Records (EHRs) based on semi-structured interviews conducted with EHR users across multiple Neonatal Intensive Care Units.

 Winner of the national AMIA Student Design Challenge (UVaToday · PRWeb)

Awards

Jefferson Cup for Outstanding Undergraduate Service · UVa Dean of Engineering, 2014

Louis T. Rader Chairperson's Award · UVa ECE, 2014

Louis T. Rader Outstanding Educator Award · UVa CS, 2014

Outstanding Teaching Assistant · UVa ECE, 2014

Winner of TEDxUVa Student Speaker Competition · TEDxUVa, 2014

Winner of AMIA Student Design Challenge · AMIA, 2013

2nd Place: Hearst Health Challenge · Startup Pitch Competition, Hearst, 2013

Winner of Amazon Social Networking Hackathon · Amazon, 2013

Winner of Amazon Web Development Hackathon · Amazon, 2013

Winner of Amazon Intern Coding Competition · Amazon, 2013

Teaching

University of Virginia

Graphics and Data Visualization

CS1501 Fall 2013

CS1501 is a course I developed and taught for the UVa School of Engineering and Applied Science. The course covered data visualization fundamentals for engineering undergraduates.

CS1501 concluded in December of 2013 with 60 students enrolled locally and hundreds more enrolled online utilizing my course materials posted on the class website. The class was included in Exploring the use of student taught classes to introduce new technical topics to engineering undergraduates by Simhal et al. and presented at the IEEE Frontiers in Education Conference in 2014.

Machine Vision

ECE4502 Spring 2013

ECE4502 is a project-oriented student-led class on embedded computer vision I developed with professors Joanne Dugan and Harry Powell. The class explored embedded systems, computer vision, robotics, and rapid prototyping to build a rapid gumball sorter. (7 students)

Embedded Systems (T.A.)

ECE3440 Fall 2012 - Spring 2014

I led laboratory exercises, held office hours, graded, and occasionally lectured for UVa's undergraduate Embedded Systems course. (4 semesters; 167 students total)

Intro to Programming (T.A.)

ECE3440 Spring 2012

I led laboratory exercises and graded for UVa's undergraduate Intro to Programming course. (45 students)

Bandit Theater

Education Lead

Dec. 2018 - Present

Bandit Theater is a nonprofit improv and sketch comedy theater based in Seattle. As Bandit's volunteer Education Lead I develop curricula, teach many classes on comedy performance and public speaking, and lead corporate teambuilding exercises. In short—I organize our comedy school.

Improv comedy is a unique and supportive team-based artform that, I believe, has a place for everyone. I delight in using comedy to help others develop their creative voices.

FRC Team 6503: Iron Dragon

Mentor

Oct 2017 - Mar 2020

Through Seattle World School, I taught Java programming, basic electronics, and laser cutting to a team of international and multilingual students competing in FIRST Robotics Competition. These young engineers completed amazing robotics projects on a small team facing unique language challenges. I could not be more proud of them!

TEDx Talk

How I Accidentally Ruled A World Power (And Nearly Paid The Price For It)

TEDxUVA, 2014

The bizarre (and hopefully inspiring) story of The Kingdom of Ardent, a made-up country I "ruled" that rose to international fame and local notoriety. This comedic presentation points toward meaningful lessons on leadership and creativity in the digital age.

Art
Four Cranes

Four Cranes

Digital, 2022

An interactive conversation between four cranes over Seattle.

N+1-Tris

(n+1)Tris

Digital, 2021

A game of Tetris in which every cleared row increases the tiles' complexity.

World Record: Dodectris (Please email me if you ever beat it.)

Concerning The Power Lines On Bellevue Avenue

Concerning The Power Lines On Bellevue Avenue

Digital, 2021

A digital installation and interactive sound piece. (Try plucking the power lines.)

Another Day At Summer Camp

Another Day At Summer Camp

Digital, 2016

A digital meditation on repetition and movement.

50 Lines

50 Lines

Installation, 2015

A physical installation at Seattle Maker Faire inspired by the wall drawings of Sol Lewitt

Variations

Variations

Installation, 2015

A physical installation at Seattle Maker Faire exploring every combination of n lines in an m sided regular polygon.

Curated Snaps

Curated Snaps

Digital, 2014-2018

A gallery of experiments in modern art created on—and presented through—an ephemeral social app.

If You Were Here

If You Were Here

Digital, 2014

Funded by the Jefferson Trust, this project documents the soundscape architecture of the Academical Village at UVa by simulating the reverb of various locations on input audio samples.

Snowballs On Steps (For David Hammons)

Snowballs On Steps (For David Hammons)

Installation, 2013

A collaborative installation with Sanda Iliescu at the UVa School of Architecture in homage to Bliz-aard Balls by David Hammons.

Painting the American Flag and Works on Paper

Painting the American Flag & Works on Paper

Painting, 2013

A group exhibition of paintings at the Elmaleh Gallery at the University of Virginia.

A Piece In D-Sharp Minor

A Piece in D-Sharp Minor

Audio, 2012

A piece of experimental music composed in the browser using simultaneous YouTube videos.

Un-painting: From Yellow to White

Un-painting: From Yellow to White

Painting, 2011

An ephemeral group installation in which random number generators were used to paint the entryway at the UVa School of Architecture.