Open-Model Safety & Evaluation Lead

Reflection; New York (NY); On-site; USD 150,000 - 190,000; Full time

Benefits offered by this job

Top-tier compensation

Stock options

Health & wellness

Meals provided

Parental leave 22 weeks

Unlimited vacation

Visa sponsorship

Team off-sites

Job summary

Reflection is building a foundational role in model evals and safety, embedding with research and infrastructure teams to accelerate frontier model development. You’ll define evaluation frameworks, build operational infrastructure for safety, and connect evals to the model lifecycle as a 0-to-1 initiative.

You bring a first-responder mindset to cut through noise, align stakeholders, and drive resolution when things go sideways, helping Reflection shape open-weight AI responsibly.

Qualifications

• 7+ years of experience in technical program management, research operations, or ML engineering.
• Familiar with the landscape of model evaluation, AI safety, red-teaming, and the evolving regulatory and industry safety ecosystem.
• Deep technically to engage with researchers and engineers on model behavior, eval design, data pipelines, and safety-critical systems.
• Proven ability to build structures where none exists with clear ownership and durable processes.
• Strong stakeholder management across deeply technical ICs, research leadership, and external partners.
• Excited to build from zero to one and define model safety and evaluation at Reflection.

Responsibilities

• Build the foundational infrastructure for model evals and safety, defining evaluation frameworks, tooling, and processes.
• Stand up model safety operations including workflows, cadences, and decision frameworks linked to development lifecycle.
• Partner with research and engineering to embed safety checkpoints into the development process.
• Drive scoping and prioritization of eval science and infrastructure investments; decide what to build in-house vs. adopt.
• Establish Reflection's engagement with external safety ecosystem and represent safety posture to stakeholders.
• Create visibility and reporting of model safety status, evaluation coverage, and open risks for leadership.
• Champion blameless post-mortems and continuous learning to improve systems.

Skills

Technical program management

Research operations

ML engineering

Stakeholder management

Cross-team collaboration

Job description

Reflection is building a foundational role in model evals and safety, embedding with research and infrastructure teams to accelerate frontier model development. You’ll define evaluation frameworks, build operational infrastructure for safety, and connect evals to the model lifecycle as a 0-to-1 initiative.

You bring a first-responder mindset to cut through noise, align stakeholders, and drive resolution when things go sideways, helping Reflection shape open-weight AI responsibly.