Margaryta Tak

AI Data Quality and Research professional with experience evaluating multimodal AI outputs, auditing structured datasets, and applying safety, ethical, and quality standards across AI and clinical research environments. Skilled in identifying ambiguities, edge cases, and inconsistencies, documenting findings, and maintaining accuracy across high-volume review workflows. Holds a Master of Science in Psychology with training in research methodology, behavioral data analysis, and evidence-based decision-making.

AI Writing Evaluator Freelancer - Outlier AI - Remote

(2024-02)

• Evaluate generative AI model outputs across text, audio, and image modalities using established standards and structured evaluation guidelines.
• Review outputs for accuracy, consistency, completeness, quality, and adherence to defined requirements across high-volume evaluation workflows.
• Identify inconsistencies, ambiguities, edge cases, and gaps, documenting findings with attention to detail for review and resolution.
• Perform quality-control reviews and data quality assurance to improve the precision and reliability of AI training data.
• Conduct peer-output reviews to promote consistent application of quality standards across high-volume workflows.
• Perform transcription, annotation, and timecoding of audio data within structured AI data-review workflows.
• Generate clear written evaluations and summaries in English, aligned with project requirements and model-training objectives.
• Evaluate outputs for compliance with AI safety, trust and safety, ethical, and policy standards.
• Apply structured review criteria consistently while working independently across large volumes of data.

Data Labeling Analyst II - Tundra Technical Solutions (on assignment at Meta) - Remote

(2025-08 - 2026-02)

• Audited large-scale labeled datasets supporting AI training pipelines for accuracy, consistency, and quality using structured data review practices.
• Used internal data and labeling tools to monitor, validate, and improve dataset quality within structured workflows.
• Investigated data anomalies, ambiguous cases, and tool limitations, documenting detailed findings for escalation.
• Applied established data-quality guidelines and project-specific benchmarks consistently across datasets.
• Identified recurring quality issues and documented observations to support data-quality improvements.
• Maintained consistent annotation and review practices within high-volume AI data workflows.

Clinical Research Coordinator - Professional Psychiatric Services - Ohio

(2022-09 - 2024-01)

• Collected, validated, and managed structured research datasets within electronic data systems.
• Reviewed research data for discrepancies and initiated corrective actions to maintain data accuracy and integrity.
• Conducted data analysis to support research outcomes and improve process efficiency.
• Documented research workflows and maintained accurate records in accordance with established regulatory standards.
• Applied standardized research procedures consistently across data collection and documentation activities.
• Maintained organized research records requiring strong attention to detail, accuracy, and adherence to established protocols.
• Supported research activities involving structured information, data validation, documentation, and quality control.

Registered Behavior Technician - Apollo Behavior - Georgia

(2021-03 - 2021-08)

• Collected and analyzed structured behavioral data using standardized protocols.
• Maintained detailed documentation and monitored data for accuracy and consistency.
• Applied established data-collection procedures to support reliable behavioral assessment.

Master of Science - Psychology - University of Louisiana at Monroe (2020-12 - 2022-08)

Bachelor of Arts - Psychology - University of California, Davis (2017-09 - 2020-09)