Amazon

Note that the info in this review also applies for the 'AI Content Expert' role (which is similar but one level up from the ML DA role).

This role is promoted as an exciting opportunity to work with engineers and gain exposure to cutting-edge AI and LLM technologies. In reality, it’s a highly repetitive, production-line style job focused almost entirely on completing a constant queue of annotation, editing, or rating tasks. If you enjoy doing the same type of task for hours at a time under close monitoring, this could suit you. Anyone expecting variety, growth, or collaboration will likely find it frustrating and draining.

Pros:

• Reliable workload (you’ll rarely run out of tasks).
• You do get a sense of how large-scale data annotation operations support AI training.
• Decent equipment provided (company laptop, basic office setup) & the office is nice.
• Possibly a foot in the door at a big tech company, if you’re willing to endure monotonous work for a long stretch.
• The team members are incredible, the best bunch of people you'll likely work with. They are the most diverse and clever group of people, really making the job bearable.

The job is extremely repetitive and production-driven. You’re assigned tasks from a queue and complete them until the queue changes or runs out — that’s the whole job.

Every minute of your day is tracked, including breaks and how long you spend on each task. The system marks you “idle” after only a few minutes away from your screen, and you're expected to extend your working day to account for any breaks over the allowance (which is already quite limited).

No phones, watches, or earphones are allowed.

Laptops are literally secured to desks, so there’s no movement around the office.

Very limited learning or development opportunities. “Training time” only happens if you’ve completely exhausted all tasks (which basically never happens). Even then, you have to ask for approval that often isn’t guaranteed.

The work environment can feel oppressive and isolating. When you raise concerns about the monotony or unhealthy setup, the response is usually “that’s just how ops is.”

Management prioritizes throughput above all else. Efficiency metrics matter far more than people’s well-being or career development.

If you’re on a fixed-term contract, be prepared that it may simply end without renewal.

Movement into other departments/positions seems unnecessarily difficult. I feel like it's encouraged in the company in general, but for this role? Not so much.

Be honest about what this job actually is. Stop marketing it as a gateway to AI engineering or cutting-edge work when it’s essentially a repetitive data annotation role.

Introduce genuine L&D time, trust your employees more, and loosen the micromanagement. The current setup might be efficient in the short term, but it burns people out quickly.