Arabic-English NER & Text Classification for LLM Fine-Tuning
Arabic and English. My tasks included named entity recognition (NER), sentiment and topic classification, and evaluation of AI-generated responses for grammatical correctness, factual accuracy, and tone appropriateness. I also contributed to fine-tuning efforts by generating prompts and evaluating completions using RLHF techniques. Additionally, I annotated mathematical problem-solving steps to support math reasoning models. All work followed strict quality guidelines and peer-reviewed QA workflows.