Agentic Engineer

Agentic Engineer

Twinnin · London / Remote-friendly (UK hours) · Full-time

Reports to: Lan Prijatelj, CTO & Co-founder

About Twinnin

Twinnin is the registry where verified humans license their face for AI use, to brands, broadcasters, and AI labs. Verify once, set your terms, take 75% every time your likeness is used. We’re the consent, licensing, and data layer for human likeness in AI: 2,500+ verified twins across three continents, built for the EU AI Act, California §927, SAG-AFTRA, and the NO FAKES Act.

Backed by Google, NVIDIA, and SFC Capital. Covered by Deadline, IMDb, and Yahoo Finance. Small team, real revenue, moving fast.

The role

Lan has built every layer of the platform himself. The ID Wallet, the rights engine, the Polygon anchor, the Verify API. That’s no longer a one-person job. We need a second engineer who can take real ownership across the stack and ship alongside him, not wait to be handed tickets.

This is not a “support” role in the keep-the-lights-on sense. You’re the person Lan hands a hard problem to and trusts to come back with it solved. You’ll own features end to end, from schema to API to the bit that’s live in production.

How we work

Our operating principle is 80/20: 80% AI, 20% human judgment, for 200% output. Everyone here codes with agents. Lan builds with Claude Code and Cursor daily. Our non-technical founder rebuilt the company website end to end with Claude Code. If you treat AI coding tools as a crutch you’re embarrassed about, this is the wrong place. If you treat them as the way you 5x yourself and then bring taste to the 20% that matters, we’ll get on.

We don’t gatekeep code. We don’t hire ego. We ship, then we make it better.

What you’ll do

• Own features across the stack, including identity, rights, provenance, the Verify API from design through to production.

• Build and harden the rights and licensing engine: consent capture, scoped licences, revocation with a defined SLA, full lifecycle logging.

• Work on the provenance pipeline: hashes, signatures, and timestamps anchored on-chain, with no PII ever leaving encrypted off-chain storage.

• Extend the Verify API that brands, broadcasters, and AI labs integrate against selective disclosure, zero-knowledge proofs, and clean docs.

• Ship fast with AI agents, then take ownership of the quality of what ships.

• Help set engineering practice as the team grows. You’re early enough to shape how we build, not just what.

What we’re looking for

• Strong, demonstrable engineering ability. You can architect and build, not just modify.

• Genuinely agentic working style: you reach for Claude Code, Cursor, or equivalent by default and have shipped real, non-trivial work with them.

• Comfortable across backend and infrastructure. Our stack leans on Google Cloud, Python, Postgres, and Polygon (commitments only: hashes, signatures, timestamps; no PII on-chain).

• Bias to ownership. You close problems end to end and don’t need hand-holding.

• Care about correctness where it counts: identity, consent, and money in, money out, leave no room for sloppiness.

• Ownership over mistakes. We are moving fast and making mistakes is okay as long as we learn and don’t repeat.

Good to have

• Identity / KYC, verifiable credentials, or digital-wallet experience.

• Blockchain or cryptographic provenance work (C2PA, W3C Verifiable Credentials, ODRL).

• Exposure to regulated or compliance-heavy products.

• You’ve worked in a tiny team where the gap between you and production is short.

What you get

• Founding-level ownership and equity. You’re employee-early, and it shows in the scope and the upside.

• Direct work with the CTO as well as the co-founders. No layers, no politics.

• A category being built in real time, with regulation and demand both moving our way.

• The freedom to use the best tools without anyone telling you to do it the slow way.

The honest version

This suits someone who wants to build a category, not pad a CV. The hours are real, the standards are high, and the work is genuinely hard. Identity, rights, and provenance are unforgiving areas to get wrong. In return you get rare ownership, a front-row seat to a market that’s opening up, and a team that backs you to ship.

• To apply: Follow LinkedIn application process and email William at william@twinnin.ai (cc lan@twinnin.ai) what you’ve built and, ideally, an exported Claude Code or Cursor session you’re proud of (something where you solved a real problem end to end). We care more about that than a polished cover letter.