Vision-Language Research Engineer: Data Strategy & RL

Neura Market; New York (NY); Hybrid; USD 350,000 - 850,000; Full time

Job summary

Anthropic in San Francisco is seeking a research engineer for the Vision team to own the end-to-end data strategy for visual capabilities, including building evals, designing RL environments, and shaping reward design to scale knowledge work.

You will collaborate with pretraining, RL, and product teams, manage external data vendors, run generalization experiments, and develop QA frameworks to guard against reward hacking while ensuring quality at scale.

Qualifications

• 7+ years of ML, computer vision, and software engineering experience.
• Experience with reinforcement learning or training data design for large models.
• Familiar with vision-language systems and large-scale data pipelines.

Responsibilities

• Own the data strategy for vision capabilities end-to-end, from building evals and scaling RL environments.
• Manage technical relationships with external data vendors, including writing task specifications, evaluating visual data and annotation quality, and iterating on reward design.
• Develop and improve QA frameworks that catch reward hacking and ensure environment quality at scale.
• Run generalization experiments to measure how data strategy changes improve multimodal capabilities on held-out evaluations.
• Partner with pretraining, RL, and product teams, and do the science that shows we’re all rowing in the same direction.

Skills

ML

Computer Vision

Software Engineering

Reinforcement Learning

Data Curation

Education

Bachelor's degree in a relevant field

Job description

Anthropic in San Francisco is seeking a research engineer for the Vision team to own the end-to-end data strategy for visual capabilities, including building evals, designing RL environments, and shaping reward design to scale knowledge work.

You will collaborate with pretraining, RL, and product teams, manage external data vendors, run generalization experiments, and develop QA frameworks to guard against reward hacking while ensuring quality at scale.