AI Safety Data & Evaluation Lead

Lila Sciences, Inc.; Cambridge, San Francisco (MA, CA); Hybrid; USD 88,000 - 144,000; Full time

Benefits offered by this job

Medical, dental, vision

Paid parental leave

Commuter benefits

Lunch program

Job summary

Lila Sciences, Inc. is building a high-agency AI Safety Data and Evaluation Program in Cambridge, MA and other locations.

The role focuses on turning safety questions into executable programs, coordinating experts, and ensuring data quality to guide leadership decisions. You will run safety-focused evaluations, manage contributor onboarding, and shape the evidence base used to mitigate risks of advanced AI systems while maintaining scientific discovery capabilities.

Qualifications

• Experience delivering expert-data, research-operations or evaluation programs.
• Strong working understanding of AI safety, threat modeling and red-teaming.
• Ability to interpret safety briefs, identify data-quality risks and escalate when needed.
• Track record of setting quality standards and improving data processes.

Responsibilities

• Translate scientific briefs into executable plans with milestones and quality standards.
• Coordinate external experts and internal colleagues for project delivery.
• Manage contributor onboarding, training and support for consistent contributions.
• Run expert-data collection, red-teaming and evaluation campaigns with progress tracking.
• Maintain records, versioned guidance and usable datasets; improve workflows.
• Identify questions requiring scientific judgment and escalate to specialists.

Skills

Expert-data programs

Safety threat models

Red teaming

Data quality

Communication

Coordination

Tools

Annotation platforms

Data provenance tools

Job description

Lila Sciences, Inc. is building a high-agency AI Safety Data and Evaluation Program in Cambridge, MA and other locations.

The role focuses on turning safety questions into executable programs, coordinating experts, and ensuring data quality to guide leadership decisions. You will run safety-focused evaluations, manage contributor onboarding, and shape the evidence base used to mitigate risks of advanced AI systems while maintaining scientific discovery capabilities.