What we are
Empirq (pronounced “em-PURK”) is a publishing platform purpose-built for AI/ML practitioners: engineers, researchers, and data scientists. Writers publish technical posts with native LaTeX math, server-rendered code, and subfield-based discovery. Readers subscribe to the writers they want to follow. Money flows directly from readers to writers via Stripe Connect, with Empirq taking 10% (or 0% for founding writers).
Today, that’s the entire product and the entire business model. The 10% platform fee is what funds Empirq; we don’t run an advertising business, sell data, or operate a “Pro tier” upsell. If any of that ever needs to change, it’ll be announced openly, with reasoning, before it takes effect, not slipped into a Terms update.
What we’re not
We’re not trying to be a general-purpose newsletter platform that happens to also support technical writing. Empirq is a vertical product for one specific audience, and we’re trying to serve that audience well rather than serve every possible audience adequately.
Empirq is bootstrapped: built with personal funds, not outside investment. That means we answer directly to readers and writers, not to a board, and not to a runway clock pushing for quarterly growth.
Who runs Empirq
Sarith Khlot — the founder — does product, engineering, design, business, and customer support. As the platform grows we’ll add people thoughtfully, prioritizing writer support and engineering. We don’t plan to grow the team faster than we can grow the writer base. The founder is an engineer building for AI/ML experts, not one of them. The team will fill in the missing specialties as needed.
We talk to writers directly, often by name, in plain email, not through a ticket queue. If you email [email protected], you’re emailing the founder.
Where we’re based
Empirq LLC is registered in the Commonwealth of Virginia, with a registered office in Richmond, VA. Banking is through Mercury. Hosting is on DigitalOcean (NYC region) and Cloudflare. Email delivery is through Resend.
The technology
The platform runs on a fairly conventional modern web stack: Next.js 15, TypeScript, PostgreSQL with Prisma, Auth.js for authentication, Stripe Connect for payments, Resend for email delivery, Cloudflare R2 for asset storage, and Inngest for background jobs. The editor is built on Tiptap (ProseMirror) with custom extensions for math, code, and embeds. Code blocks are rendered server-side with Shiki; math is rendered server-side with KaTeX.
None of this is novel. We chose conventional tools because we want to spend our innovation budget on the writer and reader experience, not on rebuilding solved problems.
Our principles
Writers are customers, not inventory.
We don’t think of writers as content suppliers feeding an algorithmic feed. They’re customers, paying us 10% (or 0%) to handle the boring infrastructure of publishing. Writers keep ownership of their content and their subscriber list. Both are theirs to take, whenever, wherever.
Readers are customers, not products.
We don’t sell reader attention to advertisers. We don’t sell reader data. The reading experience is designed for reading, not for maximizing engagement metrics that feed an ad business that doesn’t exist.
Better is better than more.
We’d rather have 500 writers publishing great technical work than 50,000 writers cross-posting from elsewhere. Quality of writing > quantity of writers. Founding writer applications are reviewed personally, in part because we’re not trying to maximize signup velocity.
Boring infrastructure, interesting content.
We use boring, well-tested infrastructure on purpose. The interesting work happens in the writing, not in the platform.
How to reach us
For general questions, partnerships, or just to say hi: [email protected].
For writer support: [email protected].
For privacy and data requests: [email protected].
For copyright and DMCA matters: [email protected].
For abuse reports: [email protected].
For legal matters: [email protected].