Research scientists develop novel architectures and techniques for self-improving AI systems. Often that involves tackling fundamental research problems in test-time optimization and reward modeling, inventing new approaches that push the boundaries of what's possible.You might be a fit if...
–You've published significant research in machine learning, particularly in areas like optimization or reinforcement learning.
–You're comfortable diving deep into theoretical problems while keeping practical applications in mind.
–You enjoy bridging the gap between research ideas and production systems.
–You're excited about building systems that genuinely learn and improve from experience.
About
True intelligence isn't built—it grows. We're pioneering autonomous systems that combine human intuition with machine perseverance. Our agents learn from every interaction, developing capabilities through experience rather than pre-programming.
We believe whole-system test-time optimization is the key to achieving domain-specific superintelligence.
We work in-person in Chicago, with a culture that values deep, focused technical work, continuous learning, and spirited discussion of ideas.