Senior Researcher - Self-Evolving Agentic Systems
Core
Advance the state of the art in self-evolving LLM-based agents by designing reinforcement learning algorithms, building scalable training environments, and translating research into systems that operate at scale.
Role type
Senior Researcher - Self-Evolving Agentic Systems
Builds
Scalable training and evaluation environments for agentic tasks (coding, tool use, web/computer use, knowledge work) and self-improving agent systems.
Domain
Artificial Intelligence / Machine Learning / Reinforcement Learning
Deliverable
production ML models
Required skills
Reinforcement learning, LLM post-training, reward modeling, building scalable AI systems, independent research, experimental design, open-source development
Preferred skills
Experience with synthetic data, model graders, reward signals, hard-example mining, active learning, data-mixture optimization, multimodal datasets, feedback loops connecting evaluation and training
Technologies
Reinforcement learning algorithms, LLMs, VLMs, reward modeling frameworks, evaluation benchmarks
Responsibilities
Design reinforcement learning methods for LLM-based agents; Build scalable training and evaluation environments; Develop rigorous benchmarks for agent capability and safety; Form hypotheses and iterate on model performance; Publish at top-tier venues and release open-source artifacts
Seniority
Senior, hands-on IC
