Lukwata Sudais Abdurahman

Key Skills

Software

AWS SageMaker

CVAT

Labelbox

Label Studio

Prodigy

Scale AI

Supervisely

V7 Labs

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Prompt Response Writing SFT

Segmentation

Text Generation

Translation Localization

Freelancer Overview

I am a multi-modal data labeler with over 3 years of experience in image, video, text, and audio annotation. My work has supported AI systems in autonomous driving, healthcare, and natural language processing. I specialize in bounding box and segmentation for computer vision, LLM evaluation for multilingual text, and audio transcription for speech recognition. I’ve worked with tools like Labelbox, Label Studio, CVAT, Prodigy, and Amazon SageMaker Ground Truth, consistently delivering high-quality annotations that meet project guidelines and improve model performance. My fluency in English and Chinese allows me to contribute to diverse datasets and multilingual AI projects.

Labeling Experience

Video Moment Retrieval for Instructional Content

Label StudioVideoSegmentationClassification

This project involved annotating instructional videos by identifying and marking temporal segments that matched natural language queries. Tasks included action recognition, classification, and writing visually grounded descriptions using Label Studio. The project followed strict quality guidelines for boundary accuracy, query specificity, and complete coverage.

This project involved annotating instructional videos by identifying and marking temporal segments that matched natural language queries. Tasks included action recognition, classification, and writing visually grounded descriptions using Label Studio. The project followed strict quality guidelines for boundary accuracy, query specificity, and complete coverage.

2025 - 2025

Audio Transcription for Speech Recognition

Aws SagemakerAudioEntity Ner ClassificationClassification

Transcribed and tagged audio clips for speech and emotion detection. Identified speaker segments and labeled emotional tone. Ensured alignment with project-specific transcription standards.

Transcribed and tagged audio clips for speech and emotion detection. Identified speaker segments and labeled emotional tone. Ensured alignment with project-specific transcription standards.

2022 - 2024

LLM Output Evaluation in English

Scale AITextClassificationText Generation

Evaluated text generation outputs for coherence, relevance, and bias. Tested prompts and classified responses for multilingual NLP models. Helped improve model performance in low-resource languages.

Evaluated text generation outputs for coherence, relevance, and bias. Tested prompts and classified responses for multilingual NLP models. Helped improve model performance in low-resource languages.

2022 - 2024

Product Image and Text Categorization

LabelboxImageClassificationText Generation

Labeled product images and descriptions for category classification and attribute tagging (e.g., color, size, brand). Used Labelbox and Scale AI to annotate both visual and textual data. Ensured consistency across thousands of entries and followed detailed guidelines to support training of recommendation and search algorithms.

Labeled product images and descriptions for category classification and attribute tagging (e.g., color, size, brand). Used Labelbox and Scale AI to annotate both visual and textual data. Ensured consistency across thousands of entries and followed detailed guidelines to support training of recommendation and search algorithms.

2023 - 2023

Sports Video Annotation

CVATVideoBounding BoxClassification

Annotated sports footage including football and volleyball matches for player tracking and action recognition. Labeled key actions such as passes, spikes, blocks, and goals across multiple frames. Used CVAT and Labelbox to apply bounding boxes and track movement over time. Ensured temporal consistency and followed detailed annotation guidelines to support training of action detection models.

Annotated sports footage including football and volleyball matches for player tracking and action recognition. Labeled key actions such as passes, spikes, blocks, and goals across multiple frames. Used CVAT and Labelbox to apply bounding boxes and track movement over time. Ensured temporal consistency and followed detailed annotation guidelines to support training of action detection models.

2023 - 2023

Show more

Education

Work History