Theonix
An AI copilot for procedural mastery. Smart glasses, a ceiling camera or a mounted phone capture a procedure; a multi-agent pipeline on Google Cloud reads it; the trainee gets guidance while their hands are still moving, and an objective score where there used to be a subjective impression.
Today that means clinical training. It went live in Rutgers Health classroom use in September 2026. My co-founder, Shaheen Timmapuri, is a practicing pediatric surgeon and runs all clinical training.
The underlying problem is not specific to medicine. It shows up anywhere a procedure has a protocol and getting it right matters.
Adorama
Six brands, more than $500M in annual revenue, and a technology organization I built from nothing across New York, Ukraine and Poland.
The platform work is finished: a fifteen-year-old .NET monolith replaced by microservices on Google Cloud, and a data stack that turned 48-hour reporting into 15-minute reporting. The current work is agentic shopping, making the catalog discoverable and buyable inside Gemini, Perplexity and the assistants after them, alongside generative AI put where it changes the job rather than decorates it: customer service, merchandising, pricing.
Twelve granted US patents
Filed during my time at Verizon, mostly on how autonomous aircraft choose networks, hand off between them, and plan routes against real-world conditions. Every one links to the public record.
- US 11230377Unmanned Aerial Vehicle Platform2022
- US 10977949Selection of Networks for Communicating with Unmanned Aerial Vehicles2021
- US 10618654Unmanned Aerial Vehicle Platform2020
- US 10394858Utilization of Third Party Networks and Third Party UAV Platforms2019
- US 10380900Information Collection and Component/Software Upgrades for UAVs2019
- US 10089890Dynamic Selection of Unmanned Aerial Vehicles2018
- US 9881022Selection of Networks for Communicating with Unmanned Aerial Vehicles2018
- US 9881021Utilization of Third Party Networks and Third Party UAV Platforms2018
- US 9875454Accommodating Mobile Destinations for Unmanned Aerial Vehicles2018
- US 9811084Identifying Unmanned Aerial Vehicles for Mission Performance2017
- US 9783293Unmanned Aerial Vehicle Platform2017
- US 9762658Dynamic Reconfigurable Client Side Beaconing2017