Laura De Souza

Key Skills

Software

Clickworker

Data Annotation Tech

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Top Task Types

Action Recognition

Audio Recording

Data Collection

Evaluation Rating

Translation Localization

Freelancer Overview

I have hands-on experience working on AI training and data labeling projects through Outlier AI, where I contributed to improving large language and multimodal models by evaluating, annotating, and refining data outputs. My work involved carefully reviewing AI-generated responses for accuracy, instruction-following, and relevance, as well as identifying errors, inconsistencies, and bias. I consistently applied detailed project guidelines to ensure high-quality annotations, demonstrating strong attention to detail and the ability to follow complex, evolving instructions. What sets me apart is my analytical mindset, as well as my medical and clinical knowledge, and the ability to provide clear, structured feedback that helps models learn more effectively. I am highly comfortable comparing multiple outputs, justifying selections with evidence-based reasoning, and maintaining consistency across large datasets. My background in healthcare and medical education further strengthens my ability to handle technical content, assess factual accuracy, and work with sensitive or high-stakes information. I work efficiently both independently and within quality-controlled workflows, making me a reliable contributor to AI training and data quality initiatives.

Labeling Experience

outlier ai text

Scale AITextRelationshipClassification

I wrote, reviewed, and refined written responses to improve large language model performance. Using a web-based annotation platform, I evaluated AI-generated text for accuracy, clarity, and instruction-following, corrected errors, and rewrote content to meet tone and formatting requirements. My work emphasized consistency, quality control, and structured feedback to support reinforcement learning and reliable language model outputs.

I wrote, reviewed, and refined written responses to improve large language model performance. Using a web-based annotation platform, I evaluated AI-generated text for accuracy, clarity, and instruction-following, corrected errors, and rewrote content to meet tone and formatting requirements. My work emphasized consistency, quality control, and structured feedback to support reinforcement learning and reliable language model outputs.

2025

outlier text to audio

Scale AIAudioAudio Recording

I evaluated AI-generated speech outputs produced from text prompts. Using a web-based annotation platform, I assessed audio quality, clarity, pronunciation, pacing, and adherence to instructions such as tone or style. I verified accuracy between the original text and the spoken output, flagged errors or unnatural speech patterns, and provided structured feedback to help improve text-to-speech and audio-language models. Text-to-speech (TTS) evaluation • Audio quality assessment • Pronunciation & fluency review • Instruction-following analysis • RLHF • Attention to detail • Multimodal AI training

I evaluated AI-generated speech outputs produced from text prompts. Using a web-based annotation platform, I assessed audio quality, clarity, pronunciation, pacing, and adherence to instructions such as tone or style. I verified accuracy between the original text and the spoken output, flagged errors or unnatural speech patterns, and provided structured feedback to help improve text-to-speech and audio-language models. Text-to-speech (TTS) evaluation • Audio quality assessment • Pronunciation & fluency review • Instruction-following analysis • RLHF • Attention to detail • Multimodal AI training

2025

outlier ai training

Scale AIVideoRelationshipQuestion Answering

During my time with Outlier AI, I worked on video annotation and evaluation projects to support multimodal AI training. Using their web-based labeling platform, I reviewed short video clips to identify key visual elements, actions, and temporal cues, and compared AI-generated interpretations to ground truth. I followed detailed guidelines to accurately label and evaluate video content, provided structured feedback on model outputs, and ensured consistency and quality across all annotations. This experience strengthened my ability to interpret dynamic visual data and contribute meaningfully to computer vision model improvement. Video annotation • Temporal labeling • Multimodal data evaluation • AI output assessment • Quality control • Attention to instruction • Visual context interpretation

During my time with Outlier AI, I worked on video annotation and evaluation projects to support multimodal AI training. Using their web-based labeling platform, I reviewed short video clips to identify key visual elements, actions, and temporal cues, and compared AI-generated interpretations to ground truth. I followed detailed guidelines to accurately label and evaluate video content, provided structured feedback on model outputs, and ensured consistency and quality across all annotations. This experience strengthened my ability to interpret dynamic visual data and contribute meaningfully to computer vision model improvement. Video annotation • Temporal labeling • Multimodal data evaluation • AI output assessment • Quality control • Attention to instruction • Visual context interpretation

2025

outlier ai

Scale AIImageClassificationQuestion Answering

I gained hands-on experience in image annotation and multimodal AI training through Outlier AI, where I worked within a proprietary web-based annotation platform to label and evaluate visual data. My responsibilities included identifying objects, attributes, and contextual elements in images, as well as assessing AI-generated interpretations for accuracy, relevance, and instruction compliance. I consistently followed detailed annotation guidelines and quality standards, contributing reliable, high-quality data to improve computer vision and multimodal AI models. Skills: Image annotation, multimodal data labeling, AI output evaluation, quality assurance, RLHF, guideline-based annotation, attention to detail

I gained hands-on experience in image annotation and multimodal AI training through Outlier AI, where I worked within a proprietary web-based annotation platform to label and evaluate visual data. My responsibilities included identifying objects, attributes, and contextual elements in images, as well as assessing AI-generated interpretations for accuracy, relevance, and instruction compliance. I consistently followed detailed annotation guidelines and quality standards, contributing reliable, high-quality data to improve computer vision and multimodal AI models. Skills: Image annotation, multimodal data labeling, AI output evaluation, quality assurance, RLHF, guideline-based annotation, attention to detail

2025

Education

Work History