Muhammad Sameer

Key Skills

Software

Clickworker

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Image

Top Label Types

Bounding Box

Classification

Freelancer Overview

I am a detail-oriented data annotator and data scientist with hands-on experience in preparing, labeling, and validating both structured and unstructured datasets for AI and machine learning workflows. My skills include data annotation, quality assurance, and compliance with detailed instructions, ensuring high accuracy and consistency in every project. I have worked extensively with Python, SQL, Pandas, and NumPy, and am adept at handling text and image data, including medical imaging. Notably, I designed and deployed a full-stack deep learning system for brain tumor detection using MRI scans, where I managed the entire pipeline from data preprocessing and augmentation to model training, interpretability, and deployment. I am comfortable working independently in remote settings and excel at communicating complex data insights to non-technical stakeholders.

Labeling Experience

Data Labeling for Autonomous Systems

CVATImageBounding BoxPolygon

This project involved the systematic annotation of large-scale video and image datasets to train object detection and semantic segmentation models for autonomous vehicles. Using CVAT, I executed complex labeling tasks to ensure the model could accurately perceive and navigate dynamic urban environments. Key Responsibilities & Technical Scope Multi-Class Labeling: Identified and annotated a diverse range of urban objects, including vehicles (cars, trucks, motorcycles), pedestrians, traffic signs, and lane markings. Annotation Techniques: 2D Bounding Boxes: For rapid object detection and localization. Polygons/Masks: Precise semantic segmentation for road boundaries and irregular shapes. Polyline Annotation: Defined lane dividers and sidewalk edges for path-planning logic. Interpolation: Leveraged CVAT’s tracking features to maintain object IDs across video frames, significantly reducing manual input time. Attribute Tagging: Assigned specific metadata to objects, such as occluded

This project involved the systematic annotation of large-scale video and image datasets to train object detection and semantic segmentation models for autonomous vehicles. Using CVAT, I executed complex labeling tasks to ensure the model could accurately perceive and navigate dynamic urban environments. Key Responsibilities & Technical Scope Multi-Class Labeling: Identified and annotated a diverse range of urban objects, including vehicles (cars, trucks, motorcycles), pedestrians, traffic signs, and lane markings. Annotation Techniques: 2D Bounding Boxes: For rapid object detection and localization. Polygons/Masks: Precise semantic segmentation for road boundaries and irregular shapes. Polyline Annotation: Defined lane dividers and sidewalk edges for path-planning logic. Interpolation: Leveraged CVAT’s tracking features to maintain object IDs across video frames, significantly reducing manual input time. Attribute Tagging: Assigned specific metadata to objects, such as occluded

2025 - 2025

Education

Work History