Skip to main content

AVIAN Patent Portfolio Filed

About the author: I'm Charles Sieg, a cloud architect and platform engineer who builds apps, services, and infrastructure for Fortune 1000 clients through Vantalect. If your organization is rethinking its software strategy in the age of AI-assisted engineering, let's talk.

Today marks a major milestone for me. This morning around 5am, after a journey of nearly 18 years, I filed the remaining 25 AVIAN patents. The original was filed in October. This filing completes one of the most comprehensive adaptive learning patent portfolios ever assembled: 573 claims across 138 distinct inventions, divided into 20 branded platform clusters. When rendered into a single PDF, the full document is over 600 pages long and has 207 diagrams. It is truly massive in scope and in size.

The Portfolio

The full list of patents is below (titles simplified for public disclosure):

ApplicationDescriptionApplication #
OriginalCore adaptive learning architecture63/906,341
AAutomated knowledge structure creation64/019,660
BIntelligent question and content generation64/019,690
CSelf-organizing knowledge structures64/020,070
DAutomatic content freshness detection64/020,198
ECross-domain knowledge transfer64/020,230
FNew-subject cold start64/020,247
GKnowledge versioning and migration64/020,260
HNew-learner onboarding64/020,288
IMulti-format evidence gathering64/020,699
JLearning behavior analytics64/020,702
KCognitive state and focus detection64/020,703
LCheating and gaming detection64/020,706
MPredictive learning path optimization64/020,708
NPersonalized curriculum planning64/022,553
OExam readiness prediction64/022,555
PAdaptive activity generation64/022,586
QScenario-based assessment64/022,587
RSmart test assembly64/022,588
SConversational tutoring64/022,589
TInterview simulation64/022,987
UExplainable recommendations64/022,994
VGroup and peer learning64/023,002
WSafety and policy enforcement64/023,012
XMulti-institution federation64/023,062
YPhysical skill acquisition64/023,069

How It Started

I originally created AccelaStudy so that I could stop carrying a large deck of Turkish flashcards around with me. When I started working on it, I barely knew how to use a Mac and had never programmed anything in Objective-C, the original language required for iOS apps. Getting it done was a slog with the tools app developers had then but, on the day the App Store launched, AccelaStudy was the very first language app available. It has been in the App Store ever since and has been downloaded over 30,000,000 times.

But I never wanted to make just a flashcard app. I wanted to create a new way of learning. I wanted a system that would adapt to the student, personalize their learning experience, and optimize their trajectory to proficiency. AVIAN does all of that and more.

From Flashcards to a Patent Family

My first attempt at patenting an idea was an improved spaced repetition algorithm to be filed as Optimized Study Method for Accelerated Memory Consolidation. I eventually decided not to pursue it and merged it into AccelaStudy directly instead. AVIAN began to take shape in another patent I drafted called The AccelaStudy Method. Most of the core ideas from that remain unchanged in AVIAN and have expanded dramatically to include cross-domain transfer intelligence, adversarial detection, cognitive state modeling, scenario-based assessment, conversational retrieval, cohort-based collaborative learning, federated multi-node deployment, embodied skill acquisition, and policy governance. Everything and the kitchen sink.

The Cost

The work it has taken to finish AVIAN has been at great cost. I've always had a strong work ethic but this has required months of 7-day work weeks, 12+ hour days. Since the beginning of February, I've pulled 3 all-nighters per week. I have had to work repeated 40-hour stretches proofing, expanding, revising, hardening, diagramming, and auditing the patent specifications and diagrams.

Diagrams and beautiful-mermaid

I created the diagrams using the popular Mermaid diagram syntax and rendered them with the open source library beautiful-mermaid. I automated the generation of all 207 diagrams. I thought that would save me time. I was wrong. Alas, while the library does a pretty good job producing diagrams, it does not do a perfect job. Not good enough to file with the USPTO, which has exacting standards. So I spent weeks working with Anthropic's Opus 4.6 debugging and adding features to beautiful-mermaid. Hours and hours of generating diagrams, reviewing them closely, and reporting back to an LLM that cannot see visual mistakes, that an arrow connecting two boxes is not quite touching one of the boxes by 6 pixels. It was painstaking and brutal. I truly almost gave up. Last Sunday was a day of despair making progress in some areas only to have regressions in others. Only last night, after hundreds of commits, did we finally achieve USPTO-ready diagrams.

More Than Paper

But the patents are more than paper. AVIAN does not exist only as documents and diagrams. It has a complete reference architecture and has been fully implemented and tested. There are over 4,000 tests of the core engine and it passes them all. I'm pleased to say that it is so efficient that it can scale to over 1 million active students on AWS infrastructure costing less than $500 a month. It is also built for energy efficiency: one of the patents contains a novel use of GPUs that requires 1/10th the energy of existing methods.

What's Next

The AVIAN patent portfolio has been filed with the USPTO provisionally but the patents themselves were written for nonprovisional examination. Prior art has been exhaustively surveyed and each application's claims are cleanly differentiated against it. The patent language is hardened and ready for examiner scrutiny. However, I will not be submitting the nonprovisional applications. Instead, the Silicon Valley law firm Wilson Sonsini will be handling the patent prosecution and securing international rights. Intellectual Ventures, in Bellevue, Washington, will be handling patent monetization and valuation.

Eighteen years ago I just wanted to stop carrying flashcards. Now AVIAN represents a fundamentally new way for humans and machines(!) to learn, adapt, and grow. Every interaction makes the system smarter. Every learner makes it better for the next one. That's the vision, and now it's protected.

Let's Build Something!

I help teams ship cloud infrastructure that actually works at scale. Whether you're modernizing a legacy platform, designing a multi-region architecture from scratch, or figuring out how AI fits into your engineering workflow, I've seen your problem before. Let me help.

Currently taking on select consulting engagements through Vantalect.