AI Data Annotator
Evaluated and annotated AI-generated responses across multiple quality dimensions — including tone, accuracy, tool use, and reasoning — to improve model performance through RLHF and SFT pipelines.
Hire this AI Trainer
Sign in or create an account to invite AI Trainers to your job.
No subject matter listed
I am experienced in AI data annotation and training data evaluation, with hands-on expertise in supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF) pipelines. My work includes assessing large language model outputs for factual accuracy, reasoning quality, safety, and alignment with user intent, using structured multi-criteria rubrics. I have contributed to enhancing model reliability by identifying hallucinations, logical inconsistencies, and policy violations. My technical skills include Python, SQL, Power BI, and key Python libraries like Pandas, Numpy, and Scikit-learn. Additionally, I have built NLP-powered systems such as a drug recommendation engine using BERT embeddings and LLMs, and developed sentiment analysis models, giving me a strong foundation in both data labeling and the practical deployment of AI models across medical, electoral, and scientific domains. I am passionate about improving AI systems through meticulous data annotation and insightful evaluation.
Evaluated and annotated AI-generated responses across multiple quality dimensions — including tone, accuracy, tool use, and reasoning — to improve model performance through RLHF and SFT pipelines.
Currently working as an AI Trainer at Outlier, contributing to the development of more reliable and accurate large language models. Responsible for evaluating AI-generated responses using structured rubrics, assessing outputs for factual accuracy, logical reasoning, tone quality, and alignment with user intent. Actively identifies hallucinations, inconsistencies, and policy violations to support RLHF and SFT pipelines, helping improve overall model performance and safety.
Worked on improving AI language models through evaluation and feedback, focusing on response quality, factual accuracy, and safety. Identified issues like hallucinations and logical errors to enhance model reliability and alignment with user intent.
Master of Science, Physics
Bachelor of Science, Physics
Data Science Intern