What it means
Also called: supervised fine-tuning.
Training a model on examples of prompts paired with good answers written or approved by people. Writing those prompt-and-answer pairs is a common task for writers and subject experts.
Used in: Writing roles and expert roles in most fields.
In open roles right now
12 of the open roles we list mention “SFT” or a word for it in the title or description, on 4 platforms (as of 23 October 2026).
Most of them on: Turing (6), SuperAnnotate (4), Appen / CrowdGen (1), Mercor (1).
Domain Expert – PythonTuring·10 countries·Flexible hours
Pay set per project15 days ago
3D & LiDAR Data Annotation AnalystAppen / CrowdGen·United States
$30–32/hr1 month ago
LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness)Mercor·Location not stated·Full-time
$100–120/hr2 months ago
Legal & Compliance Specialist – Rubric DesignSuperAnnotate·United States·Flexible hours·US bar licence needed·Law
Up to $120/hr