
Data Annotation and Curation
01 Specialized Fine Tuning (SFT)
High-quality prompt–response pairs or human-written completions. Expert annotators craft or select the “ideal” model responses to reinforce desired behaviors.
02 Reinforcement Learning with Human Feedback (RLHF)
For structured tasks or multimodal inputs, annotators tag entities, intents, points-of-view, or content categories in text, images, or video.
03 Data Curation/Collection
Acquiring or generating diverse, representative data (text, images, audio,video) based on project needs and copyright compliance.








