Model Safety & Eval Program Lead (0→1)

Visa Hunt; New York (NY); On-site; USD 180,000 - 240,000; Full time

Benefits offered by this job

Top-tier compensation

Stock options

Health & wellness

Meals

Life & family

Vacation days

Sponsorship support

Team building

Job summary

Reflection is seeking an experienced Research Program Manager to embed with research and infrastructure teams, accelerating frontier model development. This 0-to-1 role defines evaluation frameworks, builds the operational safety infrastructure, and links evals to the model lifecycle.

You bring a first‑responder mindset: when things go sideways you jump in, cut through noise, align stakeholders, and drive resolution.

Qualifications

• 7+ years of experience in technical program management, research operations, or ML engineering.
• Familiar with model evaluation and AI safety including methodologies, red-teaming, and alignment research.

Responsibilities

• Build the foundational infrastructure for model evals and safety at Reflection. Define the evaluation frameworks, tooling requirements, and operational processes.
• Stand up model safety operations as a function, including workflows, cadences, and decision frameworks.
• Partner with research and engineering leads across pre-training, mid-training, and post-training to embed safety and evaluation checkpoints.
• Drive the scoping and prioritization of eval science and eval infrastructure investments.
• Establish Reflection's engagement with the external safety ecosystem, including third-party assessments and academic partnerships.
• Create visibility and reporting structures for leadership on model safety status and open risks.
• Champion blameless post-mortems and continuous learning to turn findings into improvements.

Skills

Technical PM

Research operations

ML engineering

Stakeholder management

Zero-to-one building

Tools

Model eval frameworks

Data pipelines

Safety tooling

Job description

Reflection is seeking an experienced Research Program Manager to embed with research and infrastructure teams, accelerating frontier model development. This 0-to-1 role defines evaluation frameworks, builds the operational safety infrastructure, and links evals to the model lifecycle.

You bring a first‑responder mindset: when things go sideways you jump in, cut through noise, align stakeholders, and drive resolution.