Prompt Engineering & AI Output Evaluation (self-directed)
Repeatedly practiced prompt engineering and iterative AI tool usage to evaluate LLM outputs. Focused on checking accuracy, coherence, and instruction-following quality, aligning with RLHF-style feedback and training-data evaluation. Developed a habit of debugging and refining model interactions to improve output reliability for downstream labeling needs. • Prompt + response evaluation loops • Assessment of instruction adherence • Consistency and quality checking of generated analysis/text • Systematic improvement through iteration