Overview

Join a cutting-edge generative AI team at the center of the AI revolution, where your expertise helps build and evaluate some of the most advanced AI models in the world.

Responsibilities

• Task design and development: Design challenging, real-world ML and NLP problems drawn from your area of expertise (e.g., model training and evaluation, language understanding and generation, retrieval, applied ML pipelines) that target specific capability gaps in frontier AI models.
• Spec and golden-solution generation: Integrate problems into an agentic development environment, preparing all necessary components using Python.
• Evaluation and analysis: Evaluate target model performance on your tasks.
• Headroom identification: Identify tasks where the target model fails, and classify the nature of the failure.
• Collaborate with other experts: Work alongside fellow subject-matter experts to keep evaluations consistent and accurate.

Qualifications

• Deep, hands-on experience in machine learning and/or natural language processing, from applied industry work, research, or a graduate/PhD background in the field.
• Working proficiency in Python, applied in research, industry, or open-source work (not theoretical familiarity).
• Strong command of modern ML/NLP methods, including model training and evaluation, transformers and large language models, and standard tooling.
• Ability to engage reliably for approximately 20 hours per week.
• Past experience in AI training, model evaluation, or data annotation is preferred.
• Strong written communication and the ability to work independently and manage your own time.