Role Overview
The intern will design and develop recursive self-improvement frameworks that enable AI systems to autonomously identify weaknesses, generate improvements, conduct experiments, and validate performance gains. The role supports research on agentic reinforcement learning, test-time learning, self-evolving agents, and on-policy distillation.
Responsibilities
- Design and develop recursive self-improvement frameworks.
- Conduct research on agentic reinforcement learning, test-time learning, self-evolving agents, and on-policy distillation.
- Explore novel architectures and optimization strategies for persistent learning.
- Develop self-improving agentic systems and research harnesses.
- Investigate scalable and generalizable self-improvement mechanisms.
- Collaborate with interdisciplinary research and engineering teams.
- Publish findings in top-tier conferences and journals.
Requirements
- Currently pursuing or holding a PhD or Master's degree in Computer Science, Machine Learning, Artificial Intelligence, Mathematics, or a related field.
- Demonstrated ability to conduct independent research and develop functional AI systems.
- A strong publication record or demonstrated research experience in top-tier AI conferences such as NeurIPS, ICML, ICLR, ACL, AAAI.
- Strong programming skills in Python and experience with deep learning frameworks such as PyTorch.
- Research experience in Recursive Self-Improvement, self-evolving agents, autonomous AI research, or automated algorithm discovery is an asset.
- Experience in agentic harness design, multi-agent systems, self-evaluation, and autonomous experimentation is an asset.
- Research experience in model self-improvement, including memory mechanisms and fast/slow weight architectures, is an asset.