Kunal Panchal

Key Skills

Software

CVAT

LabelImg

Roboflow

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Top Label Types

Audio Recording

Bounding Box

Computer Programming Coding

Data Collection

Evaluation Rating

Fine Tuning

Object Detection

Polygon

Relationship

Freelancer Overview

Annotated and labeled 10,000+ images of industrial equipment using LabelImg, CVAT, and Roboflow — applying bounding box and classification annotations that generated high-quality training data for a computer vision model. Reviewed labeled datasets for accuracy, consistency, and edge cases, achieving 97%+ inter-annotator agreement and resolving ambiguous labels through iterative feedback with engineering teams. Automated image preprocessing and format conversion with Python, reducing manual preparation from 8 hours to 30 minutes. Tracked annotation progress, labeling throughput, and data quality metrics through Tableau and Power BI dashboards. Validated and standardized 500+ daily structured records using SQL and Excel, catching 95% of discrepancies before downstream processing. Documented data quality guidelines and annotation standards to ensure consistency across workflows and reliable downstream AI consumption.

Labeling Experience

Data Entry & Quality Associate (AI/Data Quality)

CVATTextEvaluation Rating

As a Data Entry & Quality Associate at Central Transport, I validated and quality-checked structured shipment records using Excel, SQL, and large language models (Claude). I leveraged Claude (LLM) to detect anomalies and inconsistencies in data, ensuring data quality before downstream processing. I documented quality guidelines and validation procedures for consistent data entry workflows. • Used an LLM to flag and resolve inaccurate records in structured shipment data across multiple sources. • Consolidated and standardized data from regional tracking sheets. • Reduced manual preparation time and improved overall data quality for AI pipeline consumption. • Ensured reliable downstream usage through thorough documentation and SOPs.

As a Data Entry & Quality Associate at Central Transport, I validated and quality-checked structured shipment records using Excel, SQL, and large language models (Claude). I leveraged Claude (LLM) to detect anomalies and inconsistencies in data, ensuring data quality before downstream processing. I documented quality guidelines and validation procedures for consistent data entry workflows. • Used an LLM to flag and resolve inaccurate records in structured shipment data across multiple sources. • Consolidated and standardized data from regional tracking sheets. • Reduced manual preparation time and improved overall data quality for AI pipeline consumption. • Ensured reliable downstream usage through thorough documentation and SOPs.

2024

Data Annotation & Analytics Intern

LabelimgImageBounding Box

During my internship at L&T Energy Hydrocarbon, I annotated and labeled over 10,000 images of industrial equipment using LabelImg, Roboflow and CVAT. I followed strict annotation guidelines to ensure high-quality training data for computer vision models. My role included annotation quality assurance, review, and resolution of ambiguous labels in collaboration with the engineering team. • Labeled and classified images using bounding boxes and category classification. • Reviewed labeled datasets for accuracy and consistency. • Achieved over 97% inter-annotator agreement rate and resolved ambiguous edge cases. • Built Tableau dashboards to track labeling throughput and data quality metrics.

During my internship at L&T Energy Hydrocarbon, I annotated and labeled over 10,000 images of industrial equipment using LabelImg, Roboflow and CVAT. I followed strict annotation guidelines to ensure high-quality training data for computer vision models. My role included annotation quality assurance, review, and resolution of ambiguous labels in collaboration with the engineering team. • Labeled and classified images using bounding boxes and category classification. • Reviewed labeled datasets for accuracy and consistency. • Achieved over 97% inter-annotator agreement rate and resolved ambiguous edge cases. • Built Tableau dashboards to track labeling throughput and data quality metrics.

2024 - 2024

Education

Work History