Jessica Carvalho De Souza

Key Skills

Software

Appen

OneForma

Toloka

Scale AI

Remotasks

Mindrift

Top Subject Matter

Search Engine Advertising

Search Engine Evaluation

Generative AI (LLMs)

Top Data Types

Text

Image

Document

Top Task Types

RLHF

Segmentation

Classification

Text Generation

Question Answering

Transcription

Evaluation/Rating

Data Collection

Freelancer Overview

Ads Quality Rater. Core strengths include Internal, Proprietary Tooling, and Remotasks. AI-training focus includes data types such as Text and labeling workflows including Evaluation, Rating, and RLHF.

Labeling Experience

Social Media Evaluator & Data Analyst / Data Collection Associate

AppenTextRLHF

At Crowdgen (by Appen), I conducted generative AI evaluation, analyzing and rating responses from Large Language Models by focusing on Helpfulness, Honesty, and Harmlessness. I performed fact-checking tasks to prevent hallucinations and toxic content. I classified sensitive text data, reviewed ads, and participated in large multimedia data collection projects. • Completed high-precision audio transcription and segmentation tasks for ASR training. • Used SRT and ADAPT platforms for social media moderation and text annotation. • Ensured safe content by identifying sensitive and policy-violating materials. • Supported integrity by verifying news and viral post credibility.

At Crowdgen (by Appen), I conducted generative AI evaluation, analyzing and rating responses from Large Language Models by focusing on Helpfulness, Honesty, and Harmlessness. I performed fact-checking tasks to prevent hallucinations and toxic content. I classified sensitive text data, reviewed ads, and participated in large multimedia data collection projects. • Completed high-precision audio transcription and segmentation tasks for ASR training. • Used SRT and ADAPT platforms for social media moderation and text annotation. • Ensured safe content by identifying sensitive and policy-violating materials. • Supported integrity by verifying news and viral post credibility.

2024 - Present

AI Trainer / Data Annotator

RemotasksTextRLHF

As an AI Trainer/Data Annotator at Outlier AI, I evaluated and rewrote responses generated by Large Language Models for accuracy, safety, and logical soundness. I created effective training prompts and ideal responses to enhance AI fluency and problem-solving. I collaborated closely with team leads to resolve edge cases and adapt to dynamic project needs. • Provided RLHF-based feedback for model alignment. • Generated high-quality SFT data in Remotasks ecosystem. • Adhered to comprehensive, changing project guidelines. • Improved the factual reliability of Generative AI outputs.

As an AI Trainer/Data Annotator at Outlier AI, I evaluated and rewrote responses generated by Large Language Models for accuracy, safety, and logical soundness. I created effective training prompts and ideal responses to enhance AI fluency and problem-solving. I collaborated closely with team leads to resolve edge cases and adapt to dynamic project needs. • Provided RLHF-based feedback for model alignment. • Generated high-quality SFT data in Remotasks ecosystem. • Adhered to comprehensive, changing project guidelines. • Improved the factual reliability of Generative AI outputs.

2024 - Present

Search Engine Evaluator

Text

As a Search Engine Evaluator with RWS Group, I examined algorithmic search results for relevance, quality, and utility using the RaterHub tool. I followed comprehensive search rating guidelines to assess web content accuracy and user intent matching. I was responsible for identifying and flagging spam and unsafe content. • Evaluated web pages, images, and videos for search quality. • Used RaterHub and managed tasks on RWS Workzone portal. • Helped improve leading search engine safety and accuracy. • Flagged hate speech and disinformation consistently.

As a Search Engine Evaluator with RWS Group, I examined algorithmic search results for relevance, quality, and utility using the RaterHub tool. I followed comprehensive search rating guidelines to assess web content accuracy and user intent matching. I was responsible for identifying and flagging spam and unsafe content. • Evaluated web pages, images, and videos for search quality. • Used RaterHub and managed tasks on RWS Workzone portal. • Helped improve leading search engine safety and accuracy. • Flagged hate speech and disinformation consistently.

2023 - Present

Ads Quality Rater

Text

As an Ads Quality Rater at Welocalize, I analyzed and rated the relevance and quality of web advertisements for a major search engine. I evaluated landing page trustworthiness and user experience, ensuring strict alignment with user search intent. My role required applying comprehensive policy guidelines to assess ad content and maintain productive metrics. • Rated and flagged misleading, offensive, or low-quality ads. • Used EWOQ platform under continually updated project guidelines. • Enhanced ad ranking algorithms with feedback. • Contributed to search advertising safety and effectiveness.

As an Ads Quality Rater at Welocalize, I analyzed and rated the relevance and quality of web advertisements for a major search engine. I evaluated landing page trustworthiness and user experience, ensuring strict alignment with user search intent. My role required applying comprehensive policy guidelines to assess ad content and maintain productive metrics. • Rated and flagged misleading, offensive, or low-quality ads. • Used EWOQ platform under continually updated project guidelines. • Enhanced ad ranking algorithms with feedback. • Contributed to search advertising safety and effectiveness.

2023 - Present

Education

Work History