RLHF Specialist

About the role

The RLHF Specialist will design and operate reinforcement‑learning‑from‑human‑feedback pipelines to improve large language models. Working remotely with a global team, you will ensure model outputs are helpful, honest, and harmless while aligning them with human values.

Key responsibilities

• Generate high‑quality preference data by ranking model responses on helpfulness, honesty, and harmlessness.
• Design multi‑turn prompts that stress‑test reasoning, safety, and bias.
• Write chain‑of‑thought explanations to train reward models.
• Collaborate with ML engineers to diagnose failure modes and fill data gaps.
• Develop and iterate annotation strategies for consistent preference scoring.
• Probe models for vulnerabilities, document findings, and suggest data‑driven interventions.
• Maintain a personal benchmark set and regularly evaluate new model versions.

Required profile

• At least 2 years of experience in data annotation, model evaluation, computational linguistics, or AI trust & safety.
• Strong understanding of reinforcement‑learning concepts such as PPO, trust regions, and reward hacking.
• Hands‑on experience fine‑tuning open‑source models (e.g., Llama 2/3, Mistral, Gemma) using LoRA/QLoRA.
• Experience with human‑in‑the‑loop workflows and annotation platforms.

Required skills

• Python programming
• Deep‑learning frameworks: PyTorch, JAX, TensorFlow
• LoRA and QLoRA fine‑tuning techniques
• Annotation tools: LabelBox, Scale AI, Snorkel
• Cloud platforms: AWS SageMaker, GCP Vertex AI