Lead Full-Stack Developer - Annotation Pipeline (MagCV+)
DocumentClassification
I built and managed an automated annotation pipeline that uses GPT-4 to extract structured fields from raw resumes. I designed the labeling schema and created a human-in-the-loop review interface, ensuring high-quality, consistent ground truth data for AI training. Semantic similarity scoring and a monitoring dashboard were implemented to continually improve data relevance and annotation quality.
• Extracted key resume fields using AI-driven annotation workflows
• Built and enforced label schemas for consistency and audit-readiness
• Developed human review workflows and access control for data quality
• Monitored annotation throughput and flagged low-confidence outputs
I built and managed an automated annotation pipeline that uses GPT-4 to extract structured fields from raw resumes. I designed the labeling schema and created a human-in-the-loop review interface, ensuring high-quality, consistent ground truth data for AI training. Semantic similarity scoring and a monitoring dashboard were implemented to continually improve data relevance and annotation quality. • Extracted key resume fields using AI-driven annotation workflows • Built and enforced label schemas for consistency and audit-readiness • Developed human review workflows and access control for data quality • Monitored annotation throughput and flagged low-confidence outputs
2025 - Present