AI Safety Data Scientist - Real-World Evaluation

Valor Capital Group; New York (NY); On-site; USD 107,000 - 147,000; Full time

Job summary

ElevenLabs in London is seeking a Data Scientist to own the datasets and evaluation workflows for AI Safety. You will convert real-world product data into reliable, structured datasets for training and evaluating models, and build scalable data and evaluation systems with safety in mind.

The role blends data science, ML, AI safety, and policy, working with safety researchers, engineers, and policy specialists to translate guidelines into practical data processes, initial labeling, and scalable

Qualifications

• Experience as a Data Scientist with real-world datasets and ML models.
• Strong grasp of dataset quality: collection, cleaning, labeling, evaluation.
• Proficiency in Python and SQL for practical data workflows.

Responsibilities

• Own safety datasets end-to-end: collection, cleaning, labeling, QC, versioning, training readiness.
• Translate safety policy into clear labeling and evaluation criteria with policy teams.
• Design and manage labeling processes, onboarding, and overseeing external contributors.
• Build evaluation workflows for models in production using real-world data to track performance.
• Develop lightweight Python/SQL pipelines to speed up data work in collaboration with ML engineers.

Skills

Python

SQL

Data analysis

ML basics

Data wrangling

Tools

Pipelines

Job description

ElevenLabs in London is seeking a Data Scientist to own the datasets and evaluation workflows for AI Safety. You will convert real-world product data into reliable, structured datasets for training and evaluating models, and build scalable data and evaluation systems with safety in mind.

The role blends data science, ML, AI safety, and policy, working with safety researchers, engineers, and policy specialists to translate guidelines into practical data processes, initial labeling, and scalable