What the work actually involves

You write technical prompts in Ukrainian at the level of your own subfield — the kind of question a synthetic chemist, virologist, or radiochemist would put to a competent colleague, not a textbook exercise. You then read what the model produces and annotate it against structured rubrics: is the science correct, is the answer useful to a real practitioner, and does it handle sensitive or dual-use material the way a responsible expert would. A third strand of the work is classification — applying written guidelines to sort prompts and conversations into categories, which sounds mechanical until you hit the ambiguous cases and have to defend a call.

The safety dimension is not decoration. Chemical, biological, and radiological/nuclear domains are where dual-use judgment matters most, and the project needs people who can tell the difference between a legitimate technical explanation and operational uplift — and who can articulate that difference in writing rather than just feeling it.

What the platform screens for

  • Verifiable PhD depth. Ongoing or completed, in chemistry, biology, or a close relative. Expect follow-ups that go a layer past your abstract — instruments, reagents, failure modes, what your controls were.
  • Genuine Ukrainian scientific register. Not conversational fluency. Can you write about, say, chromatographic separation or plasmid construction in Ukrainian without silently reverting to English or Russian terminology?
  • Written English at business level, because guidelines, rubrics, and reviewer feedback are in English.
  • Safety reasoning under pressure. Screeners probe how you'd handle a request that is scientifically sound but operationally risky, and whether you can hold a line without refusing everything.
  • Specificity over breadth. A deep subfield claim that survives interrogation beats a broad claim that collapses.

Logistics

Remote and asynchronous, with an immediate start and a commitment of roughly 7 hours per week — a genuine side engagement alongside a lab, postdoc, or faculty post. Pay observed at $48–52/hr; bands on platform projects move with task type and review outcomes and are never guaranteed. Being based in Ukraine or the wider region is a stated preference, not a gate. AI/ML background is not expected; the workflow is taught.