Do disruptive research that changes how R&D is conducted — day to day inside the organizations that run on it, and across the research community. That work runs from evaluation methodology and verification through agent trajectories to the world model they train.
What you will do
- Build the infrastructure AI does research on, at real scale.
- Work with design partners in frontier labs and R&D-heavy companies, and own whether their research measurably moves faster.
- Do the original research the system depends on: evaluation methodology, reward hacking, agent trajectories, world models.
- Ship in public — our protocol is open and the work is published.
We are looking for
- Self-driven and excited to build from the ground up.
- Rigorous research mindset, attention to detail.
- AI-native: hands-on with modern AI tools and agents.
- Background in AI research, evaluation, LLM systems, or AI infrastructure.
- Technical taste — you can tell which problems matter.
- Entrepreneurial. You want to build novel systems.
Interview
- Intro call — 30 min
- Research deep dive — 60 min
- Product round — 60 min, optional
- Offer discussion — 30 min
Compensation
$120–200K · 1–3% equity · full benefits
Also open: Founding Engineer