ReadyReport
Co-inventor and builder
Generating structured reports from video with patented technology.
The problem
Frontline teams such as police departments spend significant time turning raw video footage into written, standards-compliant reports. ReadyReport addresses that gap directly.
The approach
An AI pipeline extracts and tracks objects across video frames, models temporal patterns, detects key events, and uses large language models to draft a natural-language narrative and format it to the organization’s required template, with a human review-and-edit step before submission. The underlying method is protected by granted U.S. patent 12,394,202 B2.
The outcome
A working, patent-backed product that shortens the path from footage to a review-ready report while keeping a person in control of the final output.
Fighting deepfakes for corporations and governments.
The problem
As synthetic media becomes cheap and convincing, organizations need ways to establish whether audio and video can be trusted.
The approach
Afirmity focuses on detection, authentication and provenance for synthetic-media risk: the practical trust problem behind deepfakes, connecting technical signals to the policy decisions organizations have to make.
The outcome
A venture aimed squarely at the trust-and-provenance problem, informed by hands-on work on synthetic-media risk and governance.
Bringing polling into the era of AI.
The problem
Traditional polling struggles with cost, speed and representativeness. PilPoll explores what modern AI methods change about gathering and interpreting opinion.
The approach
Applying AI techniques to the polling workflow, from collection through interpretation, while staying honest about the methodological limits of the approach.
The outcome
An applied exploration of where AI can and cannot responsibly improve opinion research.