Gregory J Stein

As a Ph.D. student in the Robust Robotics Group at MIT's Computer Science and Artificial Intelligence Laboratory, I conduct research at the intersection of Robotics and Machine Learning. While my academic background has spanned a number of topics, including applied physics, my current focus is in the area of robotics, where I have applied my varied skills to improving robotic autonomy.

I also maintain Caches to Caches, a blog devoted to my varied interests: Machine Learning, Communication, and (occasionally) Emacs.


Ph.D. Massachusetts Institute of Technology [Expected]

Aug 2015 - Feb 2020
Graduate Research Assistant

Broadly, my research involves factoring otherwise intractable navigation problems so that robotic agents can leverage prior experience to enable more intelligent planning through unknown environments. My research was a Best Paper Finalist at the 2018 Conference on Robot Learning and the accompanying talk was awarded Best Oral Presentation.

  • Factor the Bellman Equation for planning in terms of high-level actions and observations, allowing efficient estimation of expected cost when navigating through unknown environments.
  • Generate high-quality, generic, synthetic data using unsupervised image-to-image translation for robotics tasks.
  • Working towards a doctorate in Robotics under Prof. Nicholas Roy.

S.M. Massachusetts Institute of Technology

Graduate Research Assistant
Aug 2013 - Aug 2015
GPA: 5.0/5.0

Spearhead the numerical modeling effort for the ultrafast nonlinear optics projects within the Optics and Quantum Electronics Group.

  • Master's Thesis, Modeling of Nonlinear Ultrashort Optical Pulse Propagation, supervised by Prof. Franz K√§rtner and Prof. Erich Ippen in the Department of Electrical Engineering and Computer Science.
  • Simulate and experimentally achieve high harmonic generation (HHG) with photon energies extending into the so-called water window. Use unique confining geometries to enhance HHG performance.
  • Maintain an array of high-powered laser sources, including those which are cryogenically cooled or fiber based.
  • Notable coursework: 6.631 Optics & Photonics (A); Nonlinear Optics (A+, for highest overall grade)

B.S. Cornell University

Aug 2009 - Jun 2013
GPA: 4.0/4.3

  • Awarded Summa Cum Laude from the College of Engineering
  • Earned honors in Applied & Engineering Physics for exceptional undergraduate thesis, entitled Simulation of a Few-Cycle Nonlinear Laser Pulse Compressor [supervised by Prof. Alex Gaeta].
  • Trevor R. Cuykendall Memorial Award, for outstanding teaching assistantship
  • Dean's List, all eligible semesters

Research Experience [Physics]

Quantum & Nonlinear Photonics Group, Cornell University

Undergraduate Research Assistant
Sep 2011 - Jun 2013
Department of Applied Physics
Supervisor: Prof. Alexander Gaeta

Simulated the nonlinear propagation of ultrashort laser pulses in the presence of self-phase modulation and multiphoton-ionized plasma.

  • Created numerical model for the simulation of free space ultrashort pulse propagation in the presence of a multiphoton ionizing noble gas for the purposes of designing a nonlinear pulse compressor.
  • Analyzed, through experiment and simulation, the effect of initial phase on the self-focusing dynamics of ultrashort laser pulses.

Sandia National Laboratories

Summer Student Intern
Summer 2012
Ion Beam Laboratory
Supervisor: Dr. Barney Doyle

Studied the optical properties of a novel avalanche detector, originally constructed to detect the presence of a single implanted ion.

  • Measured the two-dimensional optical response profile, in part through the design of a LabView program, of a novel avalanche detector designed for sensing single ions.
  • Theoretically analyzed the error sensitivity of traditional ion beam analysis techniques to the angle of incidence.
  • Coordinated graduate student volunteers at the 2012 Conference on the Application of Accelerators in Research and Industry; Assumed leadership role and managed behind-the-scenes operations.

Princeton Plasma Physics Laboratory

National Undergraduate Fellow
Summer 2011
Magnetic Reconnection Experiment
Supervisor: Dr. Masaaki Yamada

Processed data to analyze the electron and ion behaviors in the high-energy Hall-region of a magnetic reconnection layer.

  • Compiled and analyzed data, using MATLAB, taken from the Magnetic Reconnection Experiment in an effort to study the behavior of differently charged particles in the turbulent magnetic reconnection layer.

Additional Experience

Richards, Kibbe & Orbe LLC

Summer 2009, Summer 2010

Assisted attorneys in preparing closing documents for financial transactions.

  • Developed processes and automated programs to streamline the preparation of documents in the debt-trading department of the Firm.
  • Proposed and implemented systems to create further efficiencies in the paralegal department.

Awards & Honors

Best Paper Finalist & Best Oral Presentation

Conference on Robot Learning
Oct 2018

Finalist for Best Paper at the Conference on Robot Learning (CoRL) for my work Learning over Subgoals for Efficient Navigation of Structured, Unknown Environments.

National Defense Science and Engineering Graduate Fellowship

Department of Defense
Jan 2015 - Dec 2017

Awarded full funding from the U.S. Department of Defense reserved for exceptional students pursuing a doctoral degree.

Summa Cum Laude

College of Engineering, Cornell University
May 2013

For exemplary grade point average in Cornell University's College of Engineering.

Honors in Engineering Physics

Department of Applied & Engineering Physics, Cornell University
May 2013

For exceptional Bachelor's Thesis entitled Simulation of a Few-Cycle Nonlinear Laser Pulse Compressor.

Trevor R. Cuykendall Memorial Award

Department of Applied & Engineering Physics, Cornell University
May 2013

For outstanding teaching assistantship in Applied & Engineering Physics after helping over 30 students in a senior-level fluid mechanics course.

First Place Debater

King's College Debate Tournament
May 2009

Scored first place (with a partner) in novice policy debate out of a field of roughly two-dozen teams.