Eric Leonardis

About Eric Leonardis

I am currently working on 3D pose tracking and the neural control of musculoskeletal biomechanical models using reinforcement learning and imitation learning. My abilities at the intersection of complex systems, time series analysis, and human-computer interaction design give me a unique perspective on pipeline development and data visualization.I have had 10+ years of experience mentoring, and managing machine learning, robotics, and computational neuroscience collaborations in an academic research setting. I believe in striking a balance between qualitative and quantitative methods to provide compelling results and user insights. I have spent nearly a decade honing my skills as an instructor and scientific communicator to make complex things easy to understand.

Experience

1. Postdoctoral Fellow

   Salk Institute for Biological Studies

   Aug 2022 — Present · La Jolla, CA, US

   Advised by Talmo Pereira and Eiman Azim researching deep pose tracking in animals and humans using Social LEAP Estimates Animal Pose (SLEAP), as well as deep reinforcement learning and imitation learning for musculoskeletal biomechanical models. I have played a key role in curating large datasets and engineering 3D pose estimation pipelines. I also have worked on developing a musculoskeletal model of the mouse forelimb which we train using imitation learning and reinforcement learning. I have managed teams of data scientists to build RL pipelines in JAX, Tensorflow and Pytorch.

Education

• Hofstra University

  Bachelor of Arts, Triple Major in Psychology, History, Chinese Studies

  2010 — 2014

• UC San Diego

  Master of Science - MS, Cognitive Science

  2014 — 2017

• UC San Diego

  Ph.D. - Doctor of Philosophy (Ph.D.), Cognitive Science

  2014