[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"job-senior-data-labeling-jobs-ottawa":3},{"Ques":4,"Slug":28,"Header":29,"job_category":57},{"title":5,"content":6},"Frequently Asked Questions",[7,10,13,16,19,22,25],{"A":8,"Q":9},"Yes. The Remote Type is Remote. The Ottawa modifier reflects the target talent market and search intent, while day-to-day work is performed remotely.","Is this a remote role even though it targets Ottawa?",{"A":11,"Q":12},"Senior data labeling emphasizes training data quality ownership, handling edge cases, calibration, disagreement resolution, and QA evaluation. You will also contribute to guideline improvements that directly impact model performance improvement.","What makes these senior data labeling jobs different from entry-level labeling?",{"A":14,"Q":15},"Projects may include large language model evaluation (RLHF, prompt evaluation, rubric scoring) and\u002For computer vision annotation (bounding boxes, polygons, segmentation). Assignment depends on your strengths and current demand.","Will I work on LLM projects or computer vision annotation?",{"A":17,"Q":18},"Strong annotation guidelines compliance, careful reasoning, consistency across edge cases, QA evaluation discipline, and familiarity with RLHF or prompt evaluation workflows are the most important.","What skills are most important for success?",{"A":20,"Q":21},"This posting is for FULL_TIME. Rex.zone may also list contract or freelance roles separately depending on project needs.","Is this full-time only, or are contract\u002Ffreelance options available?",{"A":23,"Q":24},"Rex.zone supports employer types including AI labs, tech startups, annotation vendors, and BPO-style delivery teams working on NLP, content safety labeling, and LLM training pipelines.","What types of employers use Rex.zone talent?",{"A":26,"Q":27},"Quality is measured through gold set accuracy, spot-check pass rates, disagreement analysis, rework rate, and adherence to annotation guidelines. For LLM work, rubric consistency and evaluator calibration are also key.","How is quality measured in these roles?","senior-data-labeling-jobs-ottawa",{"desc":30,"title":31,"content":32},"Rex.zone is hiring for Senior Data Labeling roles supporting AI\u002FML training workflows end to end, including data labeling, RLHF, prompt evaluation, QA evaluation, and training data quality audits for large language model evaluation and computer vision annotation. You will apply annotation guidelines compliance, resolve edge cases, and improve model performance through consistent labeling decisions across NLP, named entity recognition, content safety labeling, and multimodal datasets. This is a full-time remote opportunity aligned with Ottawa talent, built for experienced annotators who can mentor peers, triage quality issues, and keep LLM training pipelines reliable from sampling to final acceptance.","Senior Data Labeling Jobs in Ottawa (Remote)",[33,36,39,42,45,48,51,54],{"h2":34,"desc":35},"Job Overview","Keyword + Job Title: Senior Data Labeling Jobs in Ottawa (Remote)\nDate Posted: 25-02-2026\nCompany: Rex.zone\nCountry: US\nRemote Type: Remote\nEmployment Type: FULL_TIME\nExperience Level: Mid-Senior\nIndustry: Technology\nJob Function: Engineering\nSkills: senior data labeling, RLHF, prompt evaluation, QA evaluation, annotation guidelines compliance, training data quality, large language model evaluation, named entity recognition, computer vision annotation, content safety labeling, LLM training pipelines\nSalary Currency: USD\nSalary Min: 63360\nSalary Max: 126720\nPay Period: YEAR",{"h2":37,"desc":38},"About Rex.zone","Rex.zone connects skilled labeling talent with teams building production AI systems across NLP, computer vision, and content safety.\nWe prioritize measurable training data quality, clear annotation guidelines, and repeatable QA evaluation so model teams can trust their datasets.\nOur projects commonly support LLM training pipelines, RLHF preference data, prompt evaluation, and evaluation datasets for model performance improvement.",{"h2":40,"desc":41},"What You Will Do","Execute senior-level data labeling across text, image, and multimodal tasks with consistent decision-making and high throughput without sacrificing accuracy.\nPerform RLHF-style preference labeling and ranking, including pairwise comparisons and rationale capture aligned to rubric-based guidelines.\nRun prompt evaluation and QA evaluation to identify failure modes, ambiguity, and policy gaps that affect large language model evaluation.\nAudit training data quality by sampling, error categorization, and root-cause analysis (guideline gaps, tooling issues, edge-case drift).\nEnsure annotation guidelines compliance by flagging unclear instructions, proposing clarifications, and updating examples for edge cases.\nSupport named entity recognition and span labeling (entities, attributes, relations) with consistent schema adherence and boundary rules.\nContribute to content safety labeling (policy categories, severity, context) to reduce false positives\u002Fnegatives in safety classifiers.\nCoordinate with project leads to resolve disagreements, calibrate annotator decisions, and stabilize inter-annotator agreement targets.\nDocument decisions, escalation notes, and acceptance criteria so downstream model training teams can interpret labels correctly.",{"h2":43,"desc":44},"Core Workflows You Will Touch","Dataset intake, sampling plans, and label taxonomy mapping for new domains and evolving model objectives.\nGold set creation, consensus labeling, and disagreement analysis to improve labeling consistency and model performance improvement.\nQuality systems: spot checks, batch-level acceptance, error rate tracking, and corrective action plans.\nEvaluation datasets: prompt sets, adversarial prompts, safety probes, and rubric-driven model outputs review.\nComputer vision annotation: bounding boxes, polygons, keypoints, segmentation masks, and attribute tagging where applicable.\nNLP labeling: classification, intent, sentiment, NER, summarization evaluation, and instruction-following evaluation.\nTooling feedback loops: UI friction reporting, shortcut recommendations, and annotation speed\u002Fquality tradeoff optimization.",{"h2":46,"desc":47},"Who You Are","You have strong experience in data labeling or data annotation with a demonstrated ability to handle complex edge cases and ambiguous policy contexts.\nYou can read and apply detailed rubrics, maintain annotation guidelines compliance, and communicate when guidelines conflict with real data.\nYou are comfortable with RLHF, prompt evaluation, and QA evaluation concepts even when tasks are non-technical in execution.\nYou can mentor peers through calibration sessions, provide actionable feedback, and raise dataset risks early.\nYou write clear notes that help reviewers, QA, and model teams understand why labels were chosen.",{"h2":49,"desc":50},"Preferred Background (Not Required)","Experience with large language model evaluation, rubric design, or preference data collection.\nExposure to content safety labeling policies and structured severity\u002Fintent frameworks.\nPrior work in annotation vendors, BPO environments, AI labs, or tech startups with fast-changing labeling specs.\nHands-on familiarity with computer vision annotation and multi-pass QA processes.\nExperience tracking training data quality metrics such as disagreement rate, rework rate, and gold accuracy.",{"h2":52,"desc":53},"Employment Details","Remote Type: Remote (must remain Remote)\nEmployment Type: FULL_TIME\nLocation Targeting: Ottawa-aligned talent (role is remote; team may include distributed stakeholders)\nCompensation Range: 63360 to 126720 USD per YEAR (based on scope, domain complexity, and quality ownership)",{"h2":55,"desc":56},"How to Apply on Rex.zone","Apply through Rex.zone with an updated resume highlighting senior data labeling experience and quality ownership.\nInclude examples of task types you have labeled (NLP, computer vision annotation, content safety labeling, RLHF, prompt evaluation).\nIf available, share metrics you influenced: training data quality improvements, reduced rework, increased inter-annotator agreement, or faster QA evaluation cycles.","AI Data Operations"]