This isn’t shift work, and it isn’t real-time clinical decision-making. It’s structured, asynchronous work designed to fit around a portfolio career.
The overall rhythm
Most clinicians working as AI Trainers fit the work into evenings, non-clinical days, and short daytime blocks between other commitments. There are usually no fixed hours — tasks come with clear expectations and deadlines, and you decide when you complete them. Communication with the wider team is asynchronous and written, rather than meeting-heavy.
Monday: reviewing the week ahead (1–2 hours)
At the start of the week, the AI Trainer logs into the project workspace and reviews new task allocations. This typically involves checking a task queue, reading updated project guidance, reviewing any clarifications posted by the delivery team, and scanning relevant Slack channels. Tasks are usually clearly scoped with the clinical scenario type, what’s being asked, expected time per task, and submission deadlines — so the clinician can plan how to spread the work across the week.
Midweek: focused training work (8–10 hours total)
Most of the week is spent on focused, independent work. This falls into three types.
Writing or refining reference answers
The clinician reviews clinical prompts and writes “gold standard” responses reflecting realistic clinical reasoning, appropriate uncertainty and caveats, safe professional tone, and alignment with real-world practice. It feels similar to explaining a case to a trainee or documenting clinical reasoning — not writing for publication.
Reviewing prompts and scenarios
Some tasks involve refining how questions or scenarios are framed, ensuring they reflect realistic clinical contexts, avoid misleading assumptions, and encourage appropriate AI behaviour. This is about shaping how the AI is asked to reason.
Working independently, without interruption
Work is completed asynchronously — no live meetings, no expectation of instant replies, no need to be “online” at set times. Clinicians often work in short, concentrated blocks when it suits them.
Ongoing communication with the team (1–2 hours total)
Throughout the week, clinicians communicate with the wider team as needed — asking clarification questions in Slack, flagging ambiguous or concerning prompts, noting edge cases, and responding to feedback from reviewers. Communication is typically written, asynchronous, professional, and low-pressure. There’s a clear escalation route if you’re unsure how to handle a task.
Later in the week: feedback and iteration (2–3 hours)
Some time is spent reviewing feedback on submitted work — seeing how reference answers were used, responding to comments from evaluators, making small revisions for consistency, and learning how guidance is being interpreted across the team. This feedback loop helps clinicians calibrate their judgement to the project’s standards, similar to audit or peer review processes in clinical work.
How responsibility is shared
AI Trainers (Clinical) aren’t working in isolation. Their work sits within a wider team including AI evaluators reviewing outputs, clinical SMEs handling complex escalations, project managers coordinating delivery, and technical teams implementing changes. The clinician’s responsibility is to apply professional judgement to their assigned tasks — not to the AI system as a whole.
What the work feels like
Clinicians often describe the work as quietly challenging and reflective — similar to teaching, supervision, or audit — and easier to fit around life than rota-based work. There’s no direct patient interaction, but the work still feels clinically meaningful.
Is this realistic alongside other work?
For many clinicians, yes. A 15-hour week might look like two or three evenings of focused work, a longer block on a non-clinical day, and short check-ins spread across the week. Exact time commitment varies by project, but the work is designed to be flexible rather than intrusive.