Freelance Data Annotator | Outlier.ai | Aether Project
Contributed to the Outlier.ai Aether Project by evaluating and ranking AI-generated text responses based on the helpfulness, honesty, and harmlessness (HHH) framework. Provided high-level annotation and detection of logical fallacies and factual inaccuracies in model outputs. Delivered detailed qualitative feedback in both English and Indonesian to enhance large language model (LLM) performance. • Performed reinforcement learning from human feedback (RLHF) using project style guides. • Applied prompt engineering to ensure contextually appropriate AI results. • Maintained strict adherence to technical documentation for high accuracy. • Used the Outlier Platform and Remotasks for all annotation tasks.