AI Research Engineer (Data Labeling & Model Fine-tuning)
ImageFine Tuning
Fine-tuned SmolVLM and Florence on domain-specific image datasets to improve scene description models. Conducted evaluation and iterative adjustments by comparing CIDEr, BLEU-4, and SPICE metrics to baseline and previous checkpoints. Developed scene description datasets and labeling schemas for semantic image understanding tasks.
• Managed data preparation and curation for model fine-tuning.
• Annotated and validated outputs for accuracy and semantic relevance.
• Leveraged custom scripts and OpenCV for both data labeling and automated checks.
• Collaborated closely with research teams to ensure annotation consistency.
Fine-tuned SmolVLM and Florence on domain-specific image datasets to improve scene description models. Conducted evaluation and iterative adjustments by comparing CIDEr, BLEU-4, and SPICE metrics to baseline and previous checkpoints. Developed scene description datasets and labeling schemas for semantic image understanding tasks. • Managed data preparation and curation for model fine-tuning. • Annotated and validated outputs for accuracy and semantic relevance. • Leveraged custom scripts and OpenCV for both data labeling and automated checks. • Collaborated closely with research teams to ensure annotation consistency.
2024 - Present