PRG 02Capabilities09 entries00 / 02

What Anthony is good at.

Nine things, each with what it means for the people who hire the lab and the evidence behind it. Anthony does the work himself. The tools behind the work are listed further down.

1. CAP 01

   Offline and on-device AI

   AI that keeps working where the data cannot leave the building and there may be no connection at all.

   EvidenceSAID, an AI interpreter for medical, legal, government, and field settings, runs on the device in 110+ languages with no connection needed.

2. CAP 02

   Computer vision and decision tools

   Cameras and footage turned into answers people can act on, not a pile of detections.

   EvidenceZoneX turns raw basketball and volleyball game film into performance intelligence using detection models and GPU inference pipelines, on an engine covered by 800+ automated tests.

3. CAP 03

   Browser extensions and new products from zero

   A first product taken from a blank page to something real people use, including tools that live inside the browser where the work already happens.

   EvidenceEnVision, a Chrome extension that flags pricing leaks in healthcare purchasing and reviews the contract behind them, went from a blank page to production solo.

4. CAP 04

   Production hardening and security

   An existing system made safe to rely on: it keeps working, the right people see the right things, and a release is checked before it ships.

   EvidenceAOS, the lab's own quality-control system, makes AI-written software prove it is finished, security and release checks included.

5. CAP 05

   Product and technical leadership

   Someone senior who owns product direction and technical decisions and stays accountable for what ships.

   EvidenceAnthony led product and operations for GoldRock Holdings' AI portfolio, four products with P&L, and built SAID and ZoneX there. Before that he led Air Force logistics for 100+ people and $40M+ in assets.

6. CAP 06

   Web3 development

   Products that connect digital and physical experiences through a blockchain.

   EvidenceAnthony led LOOT8, a Web3 product merging digital and physical experiences through blockchain, as its CEO in 2024.

7. CAP 07

   Web design and websites

   A site designed and shipped end to end, from the first Figma frame to production, with the brand and launch material to go with it.

   EvidenceMany sites shipped end to end, Figma to production, including this one, along with brand systems, demo videos, and go-to-market assets.

8. CAP 08

   Regulated and disconnected environments

   Software that holds up where the rules are strict, the stakes are real, and the network may not be there.

   EvidenceRegulated and air-gapped AI systems delivered across five clinical sites with 4,000+ daily users at peak, including live emergency use.

9. CAP 09

   Building with AI agents

   Software built with AI agents doing much of the implementation, under a quality-control plane that checks their work before it counts as done.

   EvidenceAnthony is not a classically trained engineer. He builds this way every day and writes about how it goes in public, on Medium.