I work as a Member of Technical Staff at humans&. We are building collaborative models through long interactions with realistic user simulators. Previously, I led the development of the Nemo-RL reinforcement learning library used to train frontier models at NVIDIA, and before that worked on the Deep Learning inference library (TensorRT) at NVIDIA. Before NVIDIA, I was an ML engineer at a startup (Markov Corporation). Masters in ECE from Carnegie Mellon University, focused on Machine Learning and systems design.

Technical interests:

  • Large scale reinforcement learning for LLMs
  • Systems for Learning Algorithms
  • LLMs: interpretability and in-context learning
  • Intersection of learnable machines with biology, economics, and beyond...

Thanks for reading!