11 U.S. patents across AI-driven learning systems and distributed systems, plus a 2025 international (PCT) filing on AI-generated course templates. They span his work at Apollo and Pearson. Here's what each one actually does, in plain language.
His most recent invention (2025, Pearson): an AI engine that generates adaptive course templates from real-time feedback. It pulls the user data relevant to a request, finds patterns, turns them into recommendations, generates a tailored course template, then builds and delivers learning content that fits it.
Builds and updates software on the fly, tailored to each user's device. Instead of one fixed app, it assembles the right components from a library. When the device or the available components change, it swaps in a better-fit component and rebuilds automatically.
Automatically labels content based on who the user is and their context, then generates a new version of the content carrying those smart tags, so material can be organized and surfaced relevantly without any manual tagging.
Delivers learning content in the best format for your specific device and situation. When a requested resource isn't already in the ideal format, the system converts it on demand so it's always accessible and optimized for the device.
Pause an app on one device and resume exactly where you left off on another. It records every in-app event, syncs them as a session, and replays them on the second device, so the work carries across devices without interruption.
An online-learning note tool that understands the context a note was taken in, including which activity it came from and how the learner performed. It enriches each note with metadata through semantic analysis, then uses that metadata to organize and act on notes intelligently.
Helps educators build and customize course syllabi by pulling in data services to pick the best activities for each objective. The syllabus becomes a living, metadata-rich document with versioning, permissions, and links to interactive learning objects.
Stores and manages reusable "learning activities" tied to objectives, lets different roles customize an activity before it's published, and collects usage data to recommend the right activities to the right learners.
Lets each learner chart their own route through a course. The learner sets "waypoints", markers at specific points in the material, and chooses how much detail to see at each one. The course player then delivers the material in that preferred order and depth.
A recommendation system for education: it analyzes a student's interests and goals alongside course quality and characteristics to suggest the right courses or programs. It learns from what students actually do to improve future suggestions.
A safeguard for distributed transactions spanning many services. Before starting a transaction, it checks whether any required service is likely to fail. If one is, it refuses to start at all, preventing wasted resources and messy half-completed transactions.