Matt Collins

Overview

Managed a video game studio. Wrote and prioritized backlog for 4 developers and 3 artists.

Built a custom project management system in Notion.

Worked closely with game design and UX/UI to unblock production.

Approach

When I joined the company I realized organization was lacking.

No project management software was in use.

I created a custom task management system using Notion. Along with a shared backlog, I created 'bases' for each team member with only their assigned tasks, notes and links. Automation updated properties and made calculations automatically.

I wrote clear tasks with user stories, acceptance criteria and references.

Roadmapped epics on a 6-month timeline, identified bottlenecks and prioritized features to best unblock production.

“Be the easiest person to deal with and you’d be surprised how many problems it solves” – Daniel Ek

ARBO: Arena Tactics – Game Overview – YouTube

Overview

Product “Alto” is a user friendly SPR system that helps scientists study molecular interactions to find cures and pharmaceutical drug candidates.

I led the design of the “Capture Screening” experiment analysis portal. I met with product, software and hardware teams to prioritize features. I wrote user stories, product requirement documents and iteratively designed each screen in Figma from low-fidelity prototypes to high-fidelity handoff.

Approach

First, I consulted in-house PhD experts to better understand what key metrics were important in the analysis of a capture screening experiment. Listening to the experts helped me draft initial user stories and product requirement documents.

Twice a week I led meetings that brought together specialists in each aspect of the product to determine how to optimize functionality, cost and user friendliness.

Each meeting we discussed design options that I brought to the table. For every feature I drafted 2-4 options in Figma to spark productive conversations with the group. When not pitching features or determining user flows, I developed UI prototypes.

To get direct input from customers I conducted 10+ user interviews over video calls. I learned how to generate themes and present key insights, then I translated my method into a repeatable process for the company. I also created a short video highlight reel and sent it to all teams to show employees the real customers that enjoy the company’s products.

Overview

Built a simplified VR controller for elderly users that minimizes complexity and physical strain of VR sessions. I led iterative design, prototyping and testing of the accessible VR controller.

The Capstone project was completed in a team of 3 Systems Design Engineering students under the supervision of UW professors.

Approach

First, I researched to define the problem space. Using a standard Oculus Quest controller requires fine motor control and comprehension abilities that many older adults lack, preventing them from accessing basic VR experiences.

Key pain points included lack of precision when moving Oculus thumbsticks and user strain when triggering squeeze buttons. Elderly users also demonstrated confusion when facing many input options on the Oculus controller.

Minimizing the available inputs to one movement joystick, one interaction button and two turning buttons reduced user confusion while maintaining necessary functionality for basic seated VR experiences.

Anthropometrics and optimal force metrics guided design choices. I conducted a series of in-person user tests to gather feedback on three iterations of 3D printed prototypes. A 30 page paper was written to detail the discoveries identified throughout our process and the functional high-fidelity prototype was presented to a panel of university professors.

Overview

Modelled a 1963 Aeronca Tri-Champ in SolidWorks CAD using online specifications and images of my great-great uncle’s plane.

Approach

I enrolled in engineering design courses at the University of Waterloo to learn the basics of SolidWorks CAD. I learned complex techniques such as lofting and detailing from YouTube tutorials.

Without a strong CPU the design software would crash occasionally. I learned to make frequent backups to avoid losing progress. To save time I learned tricks like mirroring across the symmetrical axis and building with reference geometry.

Made with love, passion made the long process exciting and rewarding. I was paying homage to my great-great Uncle Wilf, a wild skydiving priest who taught me the beauty in unplanned adventures.

Overview

Product “Rho” is an AI powered radiology system that uses computer vision to detect osteoporosis from Xray scans.

During my employment at 16Bit I led product management and user interface design. Working with a team of Radiologists I gathered product requirements and designed the platform’s user interface from scratch. The web app allows doctors to view scan results, manage patient information and analyze data across all of their patients.

Approach

First, I conducted research and met with radiologists to understand the radiologist profession. This helped me determine what key tools and analytics would be valuable to the doctors using product Rho.

Weekly I met with stakeholders to determine product requirements, constraints and discuss UX/UI flows that I prototyped.

Through an iterative design process I developed the product interface from low to high-fidelity and conducted user testing with prospective clients to identify pain points and areas for improvement.

Overview

Built AI object detection systems on lightweight PCBs using Python, OpenCV and TensorFlow YOLOv4.

Researched and applied cutting edge computer vision principles to optimize system detection speeds and accuracy on a modified Jetson Nano AI computer.

Approach

Minimal project oversight taught me how to efficiently use online resources to learn independently (a lot of Stack Overflow).

As the sole machine learning programmer I read articles and watched videos to learn to code sleek real-time object detection systems that run accurately without draining CPU power.

Over the 4 month term I iteratively designed and tested prototypes to produce a functional system that when mounted to a vehicle is able to detect traffic lights, stops signs, animals and vehicles in real-time with >92% accuracy.