Generative AI Engineer (Freelance), Soul AI
Contributed to generative-AI workflows by performing prompt-response analysis and rubric-based review of LLM outputs. Evaluated instruction following, reasoning-quality signals, factual consistency, and overall response usefulness, including failure-mode support for scientific reasoning. Refined outputs through documentation-heavy remote iteration focused on accuracy and clarity. • Assessed technical correctness and explanatory clarity for physics/mathematics content • Supported structured evaluation processes for model-facing use cases • Performed domain-expert validation of response content coherence • Maintained quality-oriented documentation and rapid feedback loops