Applied AI Safety & Evaluation Researcher

About the role

Join the team behind some of the world's most-loved audio personalization features, reaching millions of daily listeners. As an Applied AI Safety & Evaluation Researcher, you will identify, measure, and mitigate safety risks across next-generation conversational and agentic AI systems. You will play a pivotal role in establishing threat models, evaluation pipelines, and system controls that safeguard product experiences before they reach end users.

Responsibilities

• Develop product-specific threat models and harm taxonomies for conversational, recommender, and tool-using AI systems.
• Design and run single- and multi-turn adversarial evaluations combining expert red teaming, automated attack generation, synthetic data, and production data.
• Build reusable Python evaluation pipelines, LLM-as-a-judge workflows, regression tests, interactive dashboards, and curated golden datasets.
• Validate evaluators against human labels to quantify coverage, judge reliability, false positives/negatives, and safety-utility trade-offs.
• Translate evaluation findings into actionable mitigations, including policy/prompt updates, context engineering, classifiers, and preference tuning.
• Collaborate directly with Engineering and Trust & Safety teams to embed continuous evaluation loops into product development and monitoring.

Qualifications

• Demonstrated track record of delivering safety evaluations or mitigations for live AI/ML products.
• Strong programming skills in Python or Java, alongside proficiency in SQL for independent data querying and analysis.
• Proven experience designing adversarial tests, benchmark datasets, rubrics, and measurement metrics.

Preferred Qualifications

• Experience evaluating multi-turn agents, tool-using systems, or calibrating LLM judges.
• Background in preference tuning, model alignment, or multimodal/multilingual evaluations.
• Master's degree or PhD in Computer Science, Artificial Intelligence, Machine Learning, or a related field.

Skills

• AI
• Large Datasets
• Software Programming
• Data Analysis
• Generative AI
• Java
• Language Models
• LLM
• Python
• SQL
• AI Safety
• resourceful
• proactive
• self-driven
• automation
• calibration
• Multilingual
• Personalization
• product requirements
• Safety
• threat models

Location: Queens, New York
Job type: Contract
Salary: $80.66 – $86.66 per hour
Work hours: 8am to 5pm
Education: Bachelor's degree