Remote STEM Engineer (United States)

About the role

Rex.zone connects STEM professionals to real AI/ML production workflows, including LLM training pipelines, RLHF evaluation, data labeling, QA evaluation, prompt evaluation, named entity recognition, computer vision annotation, and content safety labeling.

Key Responsibilities

• Design, implement, and maintain data workflows that support machine learning and large language model evaluation.
• Execute RLHF-related processes including prompt evaluation, preference ranking, and rubric-based QA evaluation.
• Define and operationalize annotation guidelines compliance to improve training data quality and reduce label noise.
• Perform named entity recognition (NER) and schema validation checks; troubleshoot edge cases and ambiguous labeling.
• Support computer vision annotation programs (bounding boxes, polygons, keypoints) and audit inter-annotator agreement.
• Create metrics and dashboards for model performance improvement (accuracy, precision/recall, calibration, and error taxonomy).
• Contribute to content safety labeling and policy-driven evaluation for harmful, sensitive, and restricted content.
• Collaborate asynchronously with distributed teams; document decisions, experiments, and release notes.

Required Qualifications

• Bachelor’s degree (or higher) in a STEM field (CS, EE, Math, Stats, Physics, or related).
• Mid-Senior experience delivering engineering or applied data/ML work in production or research-adjacent environments.
• Proficiency with Python and common data tooling (pandas, NumPy) plus SQL for analysis and reporting.
• Understanding of ML evaluation concepts: ground truth construction, bias/variance, and dataset shift.
• Experience with quality assurance practices: sampling plans, audit checklists, and root-cause analysis.
• Ability to write clear documentation and follow structured rubrics for QA evaluation and labeling tasks.
• Comfort working fully remote with time-zone coordination across the United States.

Preferred Qualifications

• Exposure to NLP and LLM workflows (prompting, prompt evaluation, instruction tuning concepts).
• Experience with RLHF or human-in-the-loop evaluation pipelines.
• Computer vision annotation familiarity and tooling experience (CVAT, Labelbox, or similar).
• Knowledge of content safety labeling standards and policy frameworks.
• Experience with cloud platforms (AWS/GCP/Azure) and CI/CD or MLOps basics.
• Hands-on experience improving annotation guidelines compliance and inter-annotator agreement.

Compensation

Competitive hourly rate: $30–$50/hr (USD).