Joe McCrawMaking invisible systems understandable
My work across software engineering, product design, and writing traces a single common thread: taking systems that are hard to perceive or understand, and building tools, explanations, or experiences that make them tangible and clear.
That instinct started well before software. For over twelve years, in parallel with developing software, I worked as a high-resolution broadcast video engineer with IATSE Local 16—running switchers, recorders, and media servers as Engineer in Charge for large-scale live events (Salesforce, Oracle, Facebook, Apple, Cisco, Google). In a live broadcast, any mistake is immediately visible to thousands of people in real time. That discipline—understanding a live signal chain thoroughly enough to keep the complexity invisible to the audience—carried directly into how I build software.
My early work focused on exposing the opaque behavior of machine-vision systems through small, playful experiments (AIBias). Later product work turned invisible physical phenomena into visible tools—visualizing Wi-Fi signal coverage in 3D augmented reality (ColdSpotting) and calibrating color white points across edge-blended projectors (BlendBuddy). Through a research collaboration with Dovetail Labs and Nesta, we turned technical critiques of emotion-recognition systems into public, participatory games (Emojify). More recently, I've built tools to give engineers and AI coding agents structured, verifiable accessibility guidance (Swift Accessibility Agent, ViewLens).
My writing follows the same approach: field notes on media systems, product building, developer tools, and accessibility investigations—each piece exploring how to make complicated systems easier to understand and debug.
Focus areas
Engineering: Senior software engineer on the iOS/tvOS video playback team at Paramount+, focusing on Swift, SwiftUI, AVFoundation, and performance investigations across iOS, tvOS, and visionOS for 77M+ users. On the side, I build open-source accessibility and developer tooling.
Live production background: Twelve-plus years as a union broadcast video engineer (IATSE Local 16), running high-resolution multi-screen operations, matrix routers, and media servers for major keynotes and live broadcasts. Troubleshooting real-time signal chains under pressure remains the foundation of my engineering mindset.
Open source & research: A collection of public tools and active experiments exploring native UI inspection, accessibility training games, computer vision bias, and developer workflows. Most are available on GitHub at github.com/SerialForBreakfast.
Patents and publications: Co-inventor of U.S. Patent 12,142,000 for viewer-position optimization relative to displays—connecting computer vision, accessibility, and video viewing ergonomics. Co-author of Why Mass Unemployment?: Its Hidden Causes and Outrageous Consequences (Trafford Publishing, 2006) with Jack Stone.
Areas of interest: Making failure modes safe to practice, building evidence-first debugging workflows, turning tacit knowledge into repeatable tools, and designing systems that are honest about their limitations.