Data Scientist working on cattle re-identification (AI Training/Data Labeling)
ImageObject Detection
Worked on a biometric cattle re-identification system, requiring meticulous labeling of animal images for model training and validation. Oversaw the annotation of production datasets to enable efficient metric learning and high-accuracy identification. Validated annotated datasets through ROC-AUC and Recall@K analysis to ensure deployment readiness.
• Prepared image datasets using custom internal tooling and maintained consistency standards.
• Coordinated with the ML team to optimize labeling schemas for triplet loss models.
• Utilized similarity scoring and embedding diagnostics for label verification.
• Ensured data privacy and thorough documentation of annotation protocols.
Worked on a biometric cattle re-identification system, requiring meticulous labeling of animal images for model training and validation. Oversaw the annotation of production datasets to enable efficient metric learning and high-accuracy identification. Validated annotated datasets through ROC-AUC and Recall@K analysis to ensure deployment readiness. • Prepared image datasets using custom internal tooling and maintained consistency standards. • Coordinated with the ML team to optimize labeling schemas for triplet loss models. • Utilized similarity scoring and embedding diagnostics for label verification. • Ensured data privacy and thorough documentation of annotation protocols.
2025 - Present