AI Safety Data Scientist: Datasets & Evaluation

Elevenlabs; United States; Remote; USD 113,000 - 153,000; Full time

Benefits offered by this job

Innovative culture

Growth paths

Learning & development stipend

Social travel stipend

Annual company offsite

Co-working stipend

Job summary

ElevenLabs is seeking a Data Scientist in London to own datasets and evaluation workflows underpinning AI Safety. You will turn real-world product data into reliable data for training and evaluating models, and help build scalable data processes.

You'll collaborate with safety researchers, engineers, and policy experts to translate guidelines into practical data systems, starting with data collection, labelling, and analysis, then building scalable assessment methods.

Qualifications

• Experience as a Data Scientist or similar with real-world datasets and ML models.
• Ability to translate safety guidelines into practical data & evaluation systems.
• Strong programming and data-wrangling skills in Python/SQL.

Responsibilities

• Own safety datasets end-to-end: collection, cleaning, labelling, quality control, versioning.
• Translate safety policy into clear labelling and evaluation criteria with policy teams.
• Design and manage labelling processes with external contributors to a high quality bar.
• Build evaluation workflows for models in production using real-world data to track performance.
• Develop Python/SQL pipelines and tooling for faster, reproducible data work in collaboration with ML engineers.

Skills

Autonomy

Clear communicator

Critical thinking

Ambiguity tolerance

Education

Bachelor's degree in a relevant field

Tools

Python

SQL

Job description

ElevenLabs is seeking a Data Scientist in London to own datasets and evaluation workflows underpinning AI Safety. You will turn real-world product data into reliable data for training and evaluating models, and help build scalable data processes.

You'll collaborate with safety researchers, engineers, and policy experts to translate guidelines into practical data systems, starting with data collection, labelling, and analysis, then building scalable assessment methods.