Maaz Equbal

At Innodata India Pvt. Ltd., I annotate, validate, and quality-review AI-generated content to support machine learning model improvements. I identify hallucinations, factual inaccuracies, inconsistencies, and policy violations while meeting productivity and quality targets.

Previously, at DataAnnotation.tech, I labeled, categorized, and reviewed multilingual text datasets. I evaluated AI responses for correctness, reasoning quality, and safety, and suggested workflow improvements that increased annotation efficiency.

Through an AI Response Evaluation and RLHF project, I assessed outputs for accuracy, instruction adherence, factual correctness, and reasoning quality. I provide structured feedback across complex prompts and subject domains to support next-generation LLM training.

I'm comfortable working with annotation guidelines, dataset validation processes, QA/QC standards, and productivity-driven annotation environments. I also bring basic Python and SQL knowledge, along with strong attention to detail and analytical thinking.