[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"job-online-stem-jobs-brazil":3},{"Ques":4,"Slug":31,"Header":32,"job_category":63},{"title":5,"content":6},"Frequently Asked Questions",[7,10,13,16,19,22,25,28],{"A":8,"Q":9},"It refers to remote, full-time STEM work that can be performed online by candidates based in Brazil, focused on engineering-style AI\u002FML evaluation and LLM training pipeline support through Rex.zone.","What does “online STEM jobs in Brazil” mean for this posting?",{"A":11,"Q":12},"Yes. The role is marked Remote and is designed for distributed collaboration using online tooling and structured evaluation workflows.","Is this role strictly remote?",{"A":14,"Q":15},"You will run LLM evaluation, prompt evaluation, and QA evaluation; support RLHF preference data workflows; audit data labeling outputs; and document failure modes to drive model performance improvement.","What types of AI work will I do day to day?",{"A":17,"Q":18},"No. The role is engineering-oriented and focuses on evaluation design, QA processes, rubric construction, data quality, and reproducible measurement rather than novel model training research.","Do I need to be an ML researcher to qualify?",{"A":20,"Q":21},"Common domains include NLP tasks (classification, summarization, NER), content safety labeling and policy compliance, and sometimes computer vision annotation review for multimodal systems.","Which domains might I evaluate?",{"A":23,"Q":24},"Strong AI\u002FML evaluation fundamentals, RLHF and rubric-based grading concepts, annotation guidelines compliance, training data quality methods, and the ability to communicate clearly with stakeholders.","What skills are most important to succeed?",{"A":26,"Q":27},"This page lists a full-time remote role, but Rex.zone may also feature contract and freelance options depending on project availability.","Are contract or freelance options available?",{"A":29,"Q":30},"Quality is managed through gold sets, calibration sessions, blind audits, inter-annotator agreement checks, and continuous updates to annotation guidelines based on observed errors.","How is quality ensured across reviewers?","online-stem-jobs-brazil",{"desc":33,"title":34,"content":35},"Online STEM jobs in Brazil at Rex.zone connect engineers and applied scientists with real-world AI\u002FML training workflows, including RLHF, data labeling, QA evaluation, and prompt evaluation for large language models. You will help build training data quality, annotation guidelines compliance, and model performance improvement across NLP, computer vision, and content safety labeling. These full-time remote roles support AI labs, tech startups, and annotation vendors through scalable LLM training pipelines. Explore Rex.zone to apply, compare requirements, and join distributed teams delivering measurable evaluation outcomes.","Online STEM Jobs in Brazil",[36,39,42,45,48,51,54,57,60],{"h2":37,"desc":38},"Online STEM Jobs in Brazil — Remote AI\u002FML Evaluation Engineer","Title: Online STEM Jobs in Brazil — Remote AI\u002FML Evaluation Engineer | Date: 25-02-2026 | Company: Rexzone | Country: US | Remote Type: Remote | Employment Type: FULL_TIME | Experience Level: Mid-Senior | Industry: Technology | Job Function: Engineering | Skills: AI\u002FML evaluation, RLHF, prompt evaluation, LLM evaluation, data labeling QA, annotation guidelines compliance, training data quality, model performance improvement, NLP evaluation, content safety labeling | Salary Currency: USD | Salary Min: 63360 | Salary Max: 126720 | Pay Period: YEAR",{"h2":40,"desc":41},"About the Role","You will evaluate and improve AI system outputs by combining engineering rigor with structured human-in-the-loop feedback. Your work supports RLHF workflows, prompt evaluation, and QA evaluation for large language models, ensuring training data quality and measurable model performance improvement. You will collaborate with distributed teams to translate product requirements into clear evaluation rubrics, annotation guidelines, and scoring protocols that scale across languages and domains.",{"h2":43,"desc":44},"What You Will Do","You will design and run LLM evaluation tasks, including pairwise ranking, rubric-based grading, and error categorization; execute RLHF-style preference data collection and analysis; audit data labeling and QA evaluation results for annotation guidelines compliance; perform prompt evaluation and regression testing across model versions; build lightweight tooling, scripts, or spreadsheets to track quality metrics and reviewer calibration; coordinate with engineering and ops stakeholders to resolve systematic failure modes in NLP, content safety, and policy-sensitive outputs.",{"h2":46,"desc":47},"Core Workflows You Will Support","LLM training pipelines and evaluation harnesses; data labeling and review loops with gold sets and blind audits; named entity recognition and text classification evaluation for NLP; content safety labeling and policy compliance checks; computer vision annotation review when projects require multimodal evaluation; prompt libraries, test suites, and model release readiness criteria.",{"h2":49,"desc":50},"Required Qualifications","Mid-Senior experience in engineering, applied ML, QA evaluation, or data operations; strong understanding of LLM evaluation concepts such as rubrics, preference data, and RLHF; ability to write clear annotation guidelines and resolve disagreements through calibration; proficiency with spreadsheets and\u002For scripting for analysis (e.g., Python or SQL preferred); strong written communication for documenting issues, edge cases, and acceptance criteria.",{"h2":52,"desc":53},"Preferred Qualifications","Experience with NLP evaluation (classification, NER, summarization, reasoning); familiarity with content safety labeling, policy taxonomies, and risk-based QA; exposure to computer vision annotation or multimodal evaluation; experience working with AI labs, tech startups, BPOs, or annotation vendors; knowledge of inter-annotator agreement, sampling strategies, and evaluation set management.",{"h2":55,"desc":56},"How Success Is Measured","Improved training data quality and reduced label error rates; consistent annotation guidelines compliance across reviewers; clear, reproducible evaluation results tied to model performance improvement; faster identification of systematic failure modes through structured error taxonomy; reliable QA evaluation coverage for high-impact use cases.",{"h2":58,"desc":59},"Why Rex.zone","Rex.zone helps candidates find remote, full-time roles supporting AI\u002FML evaluation, data labeling, and RLHF workflows. You will work with globally distributed teams and contribute to production-grade LLM training pipelines that require disciplined QA evaluation and measurable outcomes.",{"h2":61,"desc":62},"Apply","Explore and apply on Rex.zone. Use your resume to highlight AI\u002FML evaluation, RLHF exposure, prompt evaluation experience, and examples of improving training data quality through QA evaluation and annotation guidelines compliance.","Engineering"]