Diana Rabemanantsoa

Key Skills

Software

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Image

Video

Audio

Top Label Types

Bounding Box

Point Key Point

Segmentation

Emotion Recognition

Transcription

Freelancer Overview

I am a dedicated data expert with hands-on experience in data entry, processing, and quality control, currently working in a fast-paced environment where accuracy and attention to detail are essential. My strengths include verifying, sorting, and annotating data to ensure high-quality outputs for validation and transcription tasks, making me well-suited for AI training data and data labeling projects. I am proficient in using internal office software tools, thrive in team settings, and consistently meet deadlines while maintaining rigorous standards. My adaptability and strong willingness to learn allow me to efficiently handle repetitive tasks and support evolving project needs in domains such as data annotation and information processing.

Labeling Experience

Voice Activity Detection

LabelboxAudioTranscription

The project focused on labeling audio data for Voice Activity Detection (VAD) to support machine learning models that identify and segment human speech from non-speech audio. The dataset is intended for applications such as speech recognition, call analytics, virtual assistants, and noise-robust audio processing systems.With annotated speech and non-speech segments within audio recordings

The project focused on labeling audio data for Voice Activity Detection (VAD) to support machine learning models that identify and segment human speech from non-speech audio. The dataset is intended for applications such as speech recognition, call analytics, virtual assistants, and noise-robust audio processing systems.With annotated speech and non-speech segments within audio recordings

2025 - 2025

Rokoko motion-capture labeling project

LabelboxVideoPoint Key PointSegmentation

The project focused on labeling and validating human motion capture data collected using Rokoko sensors to support machine learning models for animation, virtual production, and human movement analysis. The goal was to ensure accurate, consistent motion data suitable for real-time and offline use in 3D character animation pipelines. Annotated keyframes and motion sequences for full-body movements Identified and corrected tracking errors such as joint drift, jitter, occlusion, and misalignment

The project focused on labeling and validating human motion capture data collected using Rokoko sensors to support machine learning models for animation, virtual production, and human movement analysis. The goal was to ensure accurate, consistent motion data suitable for real-time and offline use in 3D character animation pipelines. Annotated keyframes and motion sequences for full-body movements Identified and corrected tracking errors such as joint drift, jitter, occlusion, and misalignment

2024 - 2025

Image annotation

LabelboxImageBounding Box

The project focused on creating high-quality labeled geospatial data for building footprint detection using high-resolution satellite and aerial imagery. The labeled dataset supports machine learning models for applications such as urban planning, mapping, disaster response, and infrastructure analysis. Annotated building footprints by drawing precise polygon boundaries around individual structures

The project focused on creating high-quality labeled geospatial data for building footprint detection using high-resolution satellite and aerial imagery. The labeled dataset supports machine learning models for applications such as urban planning, mapping, disaster response, and infrastructure analysis. Annotated building footprints by drawing precise polygon boundaries around individual structures

2024 - 2024

Education

Work History