Bekmyrza Tursyn

Key Skills

Software

AWS SageMaker

CVAT

Data Annotation Tech

Label Studio

Roboflow

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Document

Image

Text

Video

Top Label Types

Action Recognition

Bounding Box

Classification

Entity Ner Classification

Polygon

Prompt Response Writing SFT

RLHF

Segmentation

Freelancer Overview

I have extensive experience in developing and deploying AI and machine learning systems with a strong focus on data annotation, labeling, and the creation of high-quality training datasets. My work spans computer vision, NLP, and multi-agent LLM systems, where I have built and fine-tuned models for tasks such as object detection, OCR, and predictive analytics. I have engineered annotation pipelines using tools like YOLO, PyTorch, TensorFlow, and custom data labeling strategies, handling large-scale datasets—processing millions of images and thousands of annotated samples. My projects include automating real estate data workflows, defect detection in industrial settings, and building RAG architectures for NLP applications. I am skilled in optimizing annotation quality, ensuring data integrity, and integrating scalable MLOps solutions for efficient data processing and model training.

Labeling Experience

CVAT Image Labeling Project

CVATImageBounding BoxPolygon

Performed image annotation and dataset preparation for computer vision pipelines using CVAT. Covered bounding boxes, polygons, segmentation masks, and classification labels across domain-specific datasets in real estate, document AI, and visual analytics. Focused on annotation accuracy, class consistency, boundary precision, and compliance with project guidelines. Contributed to producing high-quality labeled datasets suitable for model training, validation, and production-oriented CV workflows.

Performed image annotation and dataset preparation for computer vision pipelines using CVAT. Covered bounding boxes, polygons, segmentation masks, and classification labels across domain-specific datasets in real estate, document AI, and visual analytics. Focused on annotation accuracy, class consistency, boundary precision, and compliance with project guidelines. Contributed to producing high-quality labeled datasets suitable for model training, validation, and production-oriented CV workflows.

2025 - 2025

Education

Work History