AI Research

This is a fully remote position, open to applicants anywhere in the world.

📋 Description

• Design and conduct experiments to gain insights into model behavior, test concepts, and assess datasets or training interventions.

• Employ fine-tuning and post-training experiments to confirm research hypotheses and evaluate enhancements in capabilities.

• Develop datasets, benchmarks, agent environments, evaluation rubrics, and assessment methods.

• Collaborate with AI labs and corporate partners to transform model-development objectives into research and data initiatives.

• Examine model outputs and trajectories to uncover systematic failure modes and differentiate between model, data, grader, and evaluation challenges.

• Write analytical code and create lightweight tools or prototypes, supported by engineering for dependable infrastructure.

• Communicate findings through technical reports, internal discussions, benchmarks, and scholarly papers.

• Report directly to the co-founders.

⛳️ Requirements

• Experience in conducting empirical research or similar open-ended technical projects involving machine learning systems.

• Practical experience in designing and operating agentic systems for intricate tasks, including coding agents and automated research pipelines.

• Strong experimental judgment, encompassing hypothesis development, baseline and metric selection, meticulous experimentation, and interpretation of noisy results.

• Proficiency in Python and data analysis.

• Adequate software engineering skills to prototype research workflows.

• Ability to independently convert ambiguous model inquiries into datasets, benchmarks, experiments, or other tangible contributions.

• Rigorous analytical mindset and capacity to identify subtle failure modes.

• Exceptional written and verbal communication skills with researchers, engineers, customers, and non-specialists.

• Comfort in dealing with ambiguity, rapid iteration, shifting priorities, and occasional on-site work with partner teams.

• A graduate degree, publications, lab experience, open-source contributions, independent research, and production engineering experience are beneficial indicators but not mandatory on their own.

🏝️ Benefits

• Competitive salary and stock option plan.

• Global, remote-first work environment with flexibility.

• Occasional in-person gatherings every few months.

• Direct collaboration with leading AI labs on evaluation, data, and post-training challenges.

• High level of ownership and rapid career advancement within a founding-era team.