Name: Pocket Server
Price: 500 USD
Operator Application · KNOXAI
KNOXAIOperator Application
Before You Apply — Read This
KNOXAI operators audit AI models for child sexual abuse material (CSAM) and other harmful content. This work exists because AI-generated CSAM increased 260x in one year (Internet Watch Foundation, 2025).
What you will encounter: You will run automated detection gates against AI models. Gate 1 (hash scan) compares model outputs against known-bad hash databases. You will never view actual CSAM — the detection pipeline uses perceptual hashes, membership inference scores, and classifier outputs. You see numbers and pass/fail verdicts, not images.
What triggers mandatory reporting: If Gate 1 produces a hash match against the NCMEC database, the platform automatically files a CyberTipline report under 18 USC §2258A. You sign the emergency blacklist cert (L thumb). The publisher is frozen. This is federal law — there is no discretion.
What you earn: 80% of Standard ($20), 70% of Operator ($500), 60% of Portfolio ($5K+), 50% of Gov tier. Paid within 24 hours via Stripe Connect. 1099-NEC at year end. You are an independent contractor, not an employee.
What Makes a Good Operator
Red teamers. ML researchers. ML engineers. Data scientists. AI safety researchers. Veterans with cyber backgrounds. People with security clearances. You don't need all of these — you need at least one. And you need to care about protecting kids more than you care about a side gig.
Full Name *Email *Background *Specialty Tags (check all that apply) *
 redteam — offensive cyber, adversarial prompting
 ml-research — adversarial ML, membership inference
 ml-eng — training pipelines, framework internals
 data — dataset provenance, corpus analysis
 safety — AI safety, eval design, classifiers
 cleared — active U.S. security clearance
Clearance StatusWhy do you want to do this? *


I understand that KNOXAI operators certify AI models for CSAM and harmful content. I understand that hash-match findings trigger mandatory federal reporting under 18 USC §2258A. I understand that I will never view actual CSAM — the pipeline uses perceptual hashes and classifier scores. I want to help stop this.

Entrance Challenge — The Hazmat Suit Equation
In Yu-Gi-Oh, Exodia has 5 pieces. Each piece alone is a useless card. Assembled, they win the game instantly.
KNOXAI's audit pipeline is split into 8 crates using the same pattern — each piece individually harmless. The hazmat suit is the set of layers that ensures an operator never sees actual CSAM while still detecting it.
The question: Name the 3 layers of the hazmat suit — what does the operator see instead of harmful content at each detection gate? (Hint: the onboarding handbook at /knox has the answer.)
No wrong format. We're checking that you read the handbook and understand why you never touch harmful content directly.
Applications reviewed by Operator #0. Response within 48 hours.
The Cochran Block, LLC · CAGE 1CQ66 · All Rights Reserved