Oct 2018
Dec 2025
AI Quality Evaluation Specialist
GlobalLogic Technologies Limited
Hyderabad, IN
Worked on large-scale AI evaluation and data annotation projects supporting machine learning model training and validation.
AI Quality Evaluation
- Evaluated AI model responses for correctness, grounding, personalization quality, and helpfulness.
- Conducted side-by-side (SxS) comparisons of model responses to determine the most accurate and useful output.
- Identified hallucinations, incorrect personalization, unsupported inferences, and reasoning errors in AI responses.
- Wrote clear and structured rationales explaining evaluation decisions and model rankings.
Prompt Engineering and Model Testing
- Designed multi-turn conversational prompts (15 turns) to evaluate AI reasoning and personalization capabilities.
- Tested model responses against the original prompt intent to assess accuracy and contextual understanding.
- Evaluated responses for natural integration of user context and personalization signals.
Data Annotation and AI Training
- Performed large-scale data annotation and validation tasks for AI training datasets.
- Ensured high data quality through annotation review, validation checks, and quality assurance processes.
- Maintained annotation accuracy levels above 98% across multiple AI training projects.
Japanese Language Evaluation
- Evaluated Japanese AI responses for linguistic accuracy, contextual meaning, and natural phrasing.
- Reviewed Japanese language outputs to ensure clarity, fluency, and correct interpretation of prompts.
- Assessed how effectively the AI used Japanese conversational context in personalized responses.
Analytical Evaluation
- Analyzed AI responses for grounding issues and unsupported claims.
- Verified whether model outputs correctly used conversation history and contextual information.
- Documented evaluation findings through detailed annotations and written feedback.
loading...