What the work involves

You design short multi-turn conversations — typically one to five turns — that force Gemini to draw on your actual personal context: previous chats, Gmail, Search history, YouTube activity. Then you evaluate what comes back along defined dimensions. Grounding: is every claim the model makes about you supported by real evidence, or is it a flawed inference dressed up as knowledge? Integration: is the personal detail woven in naturally, or does the model overnarrate — announcing that it checked your email before answering a simple question? Helpfulness: stripped of the personalization gimmick, is this a better response?

Most tasks end in a side-by-side comparison: two model responses, stack-ranked, with a written rationale that points to the specific turn where the problem or the win occurred. You will also extract and verify "Debug Info" to confirm which data sources the model actually pulled from, and delete evaluation conversations afterward so your test sessions don't pollute your own chat history. That housekeeping is part of the job, not an afterthought.

What the platform screens for

Turing's process runs through a Job Interest Form, a profile review, then a timed assessment that must be completed within 24 hours of being sent. The assessment is the real gate — it looks for whether you can tell the difference between a correct personalization and a lucky guess, whether your written rationale cites evidence rather than asserting a preference, and whether you can spot subtle naturalness failures in two otherwise similar responses. A BS/BA in an analytical field (policy, law, ethics, linguistics, journalism, CS) or equivalent experience is expected; prior annotation, AI evaluation, or content moderation work is strongly preferred.

Logistics

  • Fully remote, contractor engagement, roughly three months
  • At least 4 hours per day, up to 40 hours per week
  • 4 hours of daily overlap with PST required — Turing is staffing a 24-hour global operation
  • Desktop or laptop with reliable internet; mobile-only setups will not work
  • You must use your primary personal Google account, not a clean test account. This is a genuine data-access condition, not a formality — decide how you feel about it before applying.

Pay band is undisclosed on this listing. Turing typically sets rates by geography and assessed skill tier, so expect the number to surface at the pre-onboarding stage rather than in the posting.