About The Role:
This Remote, FULL_TIME AI Research position supports end-to-end research and applied experimentation for modern AI systems, including LLM training pipelines, RLHF evaluation, prompt evaluation, and model performance improvement. You will design experiments, define evaluation protocols, and improve evaluation rigor through robust measurement, dataset strategy, and QA evaluation workflows.
Key Responsibilities:
Develop and run AI research experiments across NLP, computer vision, and multimodal settings
Design evaluation harnesses for large language model evaluation, including offline/online metrics
Define RLHF and preference modeling evaluation protocols (rubrics, sampling plans, inter-rater reliability)
Partner with data labeling teams to improve training data quality and annotation guidelines compliance
Perform error analysis, ablation studies, and prompt evaluation to identify failure modes
Build reproducible pipelines for dataset curation, data labeling audits, and QA evaluation
Translate research findings into engineering-ready requirements and measurable acceptance criteria
Support content safety labeling strategy and risk analysis for model outputs
Requirements:
AI research and machine learning research fundamentals
Experience with large language models and model evaluation
Hands‑on RLHF concepts (preference data, reward modeling, evaluation)
Strong experiment design and statistics basics
Familiarity with training data quality practices and dataset documentation
NLP and/or computer vision domain knowledge
Ability to collaborate cross‑functionally with engineering and data operations
Compensation
$30–$50 per hour (HOURLY)
How To Apply:
Apply via Rex.zone with a resume and relevant research artifacts (papers, technical reports, repositories, or evaluation write-ups) demonstrating model performance improvement.
