Experience

Polycam

• United States
• Software Development
• 1 - 100 Employee

• Software Engineer, Computer Vision

  • Sep 2022 - Present

  Polycam is a phone app that allows users to rapidly and easily make 3D captures of the world around them. My role is to improve and add features to the app which are related to computer vision. Achievements: • Automatically segment & label rooms for a 2D floorplan using pre-extracted walls & objects in a few milliseconds • Automatically align images to stitch seams for 360 panoramas • Automatically detect wall colors in room mode • Automatically correct object intersections… Show more Polycam is a phone app that allows users to rapidly and easily make 3D captures of the world around them. My role is to improve and add features to the app which are related to computer vision. Achievements: • Automatically segment & label rooms for a 2D floorplan using pre-extracted walls & objects in a few milliseconds • Automatically align images to stitch seams for 360 panoramas • Automatically detect wall colors in room mode • Automatically correct object intersections in room mode • Added cylindrical crop feature (backend & part of the frontend) Show less



Magna International

• Canada
• Motor Vehicle Manufacturing
• 700 & Above Employee

• Senior Research Engineer - Computer Perception

  • Sep 2019 - Sep 2022

  Troy, MI The advanced autonomy team develops L4 self-driving capabilities. My role pertained to computer vision, with a specific focus on mapping and localization. Achievements: • Designed and coded in C++ a full SLAM algorithm from scratch to utilize a unique set of input datatypes for mapping, and a newly-invented feature type for localization. • Invented a new feature type which improves on state-of-the-art features for use in SLAM algorithms. • Improved the accuracy of the… Show more The advanced autonomy team develops L4 self-driving capabilities. My role pertained to computer vision, with a specific focus on mapping and localization. Achievements: • Designed and coded in C++ a full SLAM algorithm from scratch to utilize a unique set of input datatypes for mapping, and a newly-invented feature type for localization. • Invented a new feature type which improves on state-of-the-art features for use in SLAM algorithms. • Improved the accuracy of the state-of-the-art open-source quick line detection algorithm with less than 1% increase in computation time. • Embedded the localization portion of the SLAM algorithm using ROS2 for real-time online usage. Tested algorithm successfully in a downtown area. • Designed and coded in C++ an algorithm to quickly extract parallel lines from an image and compute the roll, pitch, and yaw of the camera assuming a Manhattan-world. • Designed and coded in C++ (integrated into ROS2) an algorithm to detect and track objects using lidar and odometry data, and resolved lidar hardware issues relating to reflectance, blockage, and ROS2 driver integration. • Designed and coded in Matlab an intrinsic (fully automated) & extrinsic camera calibration algorithm with various output formats for team requirements, including undistortion, cylindrical transformation, and bird's eye view lookup tables. Used this algorithm to calibrate cameras on 12 cars x 5 cameras each and maintain all calibrations for the team. • Developed visualization methods to analyze and improve Kalman filter performance, which was used to fuse multiple localization data streams. • Designed and coded in C++ an automated extrinsic calibration algorithm, which is now used in production, based on a prescribed custom calibration target. Show less



Stanford University

• United States
• Higher Education
• 700 & Above Employee

• PhD researcher

  • Sep 2014 - Aug 2019

  The Lentink research lab studies bird flight for inspiration into designing small aerial vehicles. My research focused on creating new techniques to quantify wing deformation of flying birds by using computer vision. Achievements: • Designed a novel automated calibration technique for multiple cameras and projectors • Created a new structured-light method to automatically reconstruct the 3D surface of flying birds at 3200 Hz • Analyzed 3D point cloud data to compute aerodynamic… Show more The Lentink research lab studies bird flight for inspiration into designing small aerial vehicles. My research focused on creating new techniques to quantify wing deformation of flying birds by using computer vision. Achievements: • Designed a novel automated calibration technique for multiple cameras and projectors • Created a new structured-light method to automatically reconstruct the 3D surface of flying birds at 3200 Hz • Analyzed 3D point cloud data to compute aerodynamic parameters of flying birds Show less



University of Illinois Urbana-Champaign

• United States
• Higher Education
• 700 & Above Employee

• Research Assistant: MechSE Rheological Fluids Laboratory

  • Aug 2012 - Aug 2014

  The Ewoldt Rheology Research Lab studies fluids like silly putty and toothpaste that can't be characterized by normal fluid dynamics equations. My research focused on mathematically characterizing splash dynamics of yield-stress fluids so we could optimally design rheological fluids for various applications including firefighting. Achievements: • Rapid prototyped droplet impact test rig & wrote Matlab program to quantify fluid dynamics automatically • Used dimensionless variables… Show more The Ewoldt Rheology Research Lab studies fluids like silly putty and toothpaste that can't be characterized by normal fluid dynamics equations. My research focused on mathematically characterizing splash dynamics of yield-stress fluids so we could optimally design rheological fluids for various applications including firefighting. Achievements: • Rapid prototyped droplet impact test rig & wrote Matlab program to quantify fluid dynamics automatically • Used dimensionless variables to compare droplet behavior across multidimensional parameter space • Measured rheology using rotational rheometer: flow curves and thixotropic time scales Show less



Chicago Transit Authority

• United States
• Truck Transportation
• 700 & Above Employee

• Engineering Intern (Infrastructure Department)

  • Jun 2012 - Aug 2012

  Reduced air compressor expenses by $18,000 annually & created a database/color-coded map in Excel to visualize the health of the Chicago subway system



Culver Academies

• United States
• Primary and Secondary Education
• 300 - 400 Employee

• Camp Counselor

  • Jun 2011 - Jul 2011

  Coordinated events for over 150 kids and was fully responsible for 11 kids at an internationally renown leadership camp



Packer Engineering Group

• Industrial Machinery Manufacturing
• 1 - 100 Employee

• STEP Intern

  • Jun 2010 - Aug 2010

  Aided in the design of a test rig, report, & technical presentation for failure analysis of concrete strands