Summary
Highly detail-oriented and analytical Data Annotator with 2+ years of experience specializing in preparing high-quality datasets for machine learning models. Proven ability to meticulously label and categorize diverse data types, including text, image, and video, adhering strictly to complex guidelines. Eager to leverage strong proficiency in data quality assurance and annotation platforms to contribute to advanced AI/ML development initiatives.
Experience
Data AnnotatorScale AI
- Annotated over 1,500 hours of audio and video data, achieving a consistent accuracy rate of 98% in labeling complex conversational nuances for NLP model training.
- Contributed to the improvement of annotation guidelines, leading to a 15% reduction in cross-annotator disagreement and enhancing dataset consistency.
Junior Data AnnotatorAppen
- Successfully completed over 25,000 text classification and entity recognition tasks for various client projects, consistently exceeding daily quotas by 20%.
- Achieved an average quality score of 96% across all projects by rigorously applying complex linguistic rules and client-specific guidelines.
Projects
Sentiment Analysis Dataset Curation
- Developed and annotated a custom dataset of 2,000 social media comments with positive, negative, and neutral sentiment labels for an academic NLP project.
- Achieved a high inter-annotator agreement rate of 0.85 (Cohen's Kappa) by creating clear, iterative annotation guidelines.
- Utilized Python scripting to preprocess raw text data, removing noise and standardizing formats before manual annotation.
Image Segmentation Annotation Guide
- Authored a comprehensive annotation guide for automotive image segmentation, detailing guidelines for labeling vehicles, pedestrians, and road signs.
- Included detailed examples and edge case scenarios, which improved new annotator onboarding efficiency by 20% and reduced initial error rates.
- Collaborated with ML engineers to ensure guide accuracy and relevance to model training objectives.
Custom Named Entity Recognition (NER) Dataset for Medical Texts
- Created a specialized dataset of 1,500 medical abstracts, annotating entities such as 'disease', 'treatment', and 'drug' to support a biomedical NLP model.
- Maintained stringent data privacy protocols while handling sensitive text, ensuring compliance with ethical guidelines.
- Conducted weekly quality checks, identifying and correcting inconsistencies to ensure a clean dataset for model training.
Education
San Jose State UniversityBachelor of Science in Cognitive Science
- Graduated Magna Cum Laude with a GPA of 3.8/4.0.
- Relevant coursework included Introduction to Artificial Intelligence, Cognitive Psychology, and Data Structures.
- Awarded Dean's Scholarship for academic excellence for 4 consecutive years.
Skills
Data Annotation & Labeling
Image SegmentationBounding Box AnnotationKeypoint AnnotationText ClassificationNamed Entity Recognition (NER)Audio Transcription
Tools & Platforms
LabelboxProdigyAmazon SageMaker Ground TruthCVATJIRAGoogle Workspace
Data Quality & Analysis
Data ValidationQuality AssuranceError IdentificationConsistency CheckingData CleansingData Interpretation
Machine Learning Concepts
Natural Language Processing (NLP)Computer VisionSupervised LearningData PreprocessingModel Evaluation (basic)
Project Management & Communication
Agile MethodologiesCross-functional CollaborationTechnical DocumentationFeedback Integration