AI moderation systems for politically sensitive platforms.
If you run a platform where every moderation decision becomes a press incident, a Congressional letter, or a user revolt, generic T&S vendors will not save you. I build the systems that do.
Senior Software Engineer at TMTG (NASDAQ: DJT), Truth Social. Five years on internal trust and safety infrastructure at scale. Founder, Eapen Technology.
Taking on 2 clients this quarter.
Apply for a Consultation01 · Services
$25,000 · 4 weeks · fixed price
Production review of your current moderation pipeline. I look at ingest, classification, queueing, reviewer tooling, appeals, and audit. You get a written architecture document, a 90-day implementation plan, and named vendor and model recommendations. Not a slide deck.
from $60,000 · 8 to 12 weeks · fixed scope
For platforms past the assessment that need the system actually built. Production-ready architecture, model evaluation harness, reviewer workflow, on-call runbook. Roadmap to first-pass automation on the categories you choose. I do the work; your team owns the result.
$15,000/month · 3-month minimum
Embedded availability for CTOs and Heads of Trust and Safety. 30 hours per month. Weekly architecture review, escalation availability, model and vendor decisions, incident post-mortems. For platforms running their own T&S build and want a senior voice in the room.
02 · About
Justus Eapen has spent the last five years as a Senior Software Engineer at Trump Media and Technology Group (NASDAQ: DJT), working on Truth Social's internal trust and safety infrastructure. He has built and operated moderation systems at scale under regulatory scrutiny, press attention, and political pressure that most platforms will never face.
Before TMTG: developer at SmartLogic, leading production systems and hosting Elixir Wizards for six seasons. Speaker at ElixirConf USA. Head of Software at Pavlok, where he scaled the engineering team to 30+. SVP of Innovation at NorthOut, running product and engineering across 4 to 6 startup teams simultaneously.
Founder of Baltimore AI/ML. Grew the community to 400+ members. Booked speakers from Johns Hopkins, NASA Goddard, and Legg Mason.
He lives in Aberdeen, Maryland with his wife and two children.
03 · Writing
Three essays I wrote for the kind of operator who would hire me. If these arguments resonate, we will probably get along.
The biggest moderation risk at a heterodox platform is not a user posting something dangerous. It is Apple or Google pulling your app. How the lean enters the policy layer, and what to do instead.
Four cases — Parler, Gab, Rumble, Telegram — and the working model of platform risk that the survivors have already internalized.
Moderation is downstream of a moral worldview. The Bay Area's worldview is incompatible with the work your platform needs done.
04 · Contact
I take on a small number of clients at a time. If you've read this far, the next step is a 30-minute call.
Apply for a Consultation