Freelance Data Annotator / LLM Trainer / Web Developer
Trained and evaluated LLMs and multimodal models using RLHF, STF, and OCR-based pipelines. Worked with complex layout and GUI-based agentic workflows to improve model behavior and task performance. • Defined and validated training/evaluation prompts for targeted outcomes. • Applied quality assurance and user testing to assess response accuracy. • Conducted red/quality checks aligned with AI safety evaluation practices. • Iterated model response optimization based on evaluation results.