Occasional thoughts on AI research, reinforcement learning, and my experiences in the field.

• Scaling RL to Non-Verifiable Domains Through Symbolic Abstractions

  RLSAF - Reinforcement Learning from Symbolic Abstraction Feedback

• Thoughts About RL and Where It Is Headed

  Reflections on reinforcement learning trends from NeurIPS 2025

• Learnings from NeurIPS 2025

  Key takeaways and insights from attending NeurIPS 2025