Reilly Raab


My resume is also available as a PDF.


I recently completed a PhD in Computer Science and Engineering at UC Santa Cruz, where my doctoral research in machine learning focused on aligning local incentives with global objectives and constraints in multiagent systems.

My previous background is in physics, with experience in scientific computing, signal processing for quantum systems, and electronics. I spent a few years between undergrad and grad school backpacking abroad and remotely developing software related to automated circuit design.

Research Interests

  • I strongly suspect that the next generation of “Artificial Intelligence” will not be focused on single models, but on multi-agent systems understood in terms of interacting units of control. In particular, does there exist a multi-agent game (for instance, based on a variational principle) such that individual agents that play the game with their environment (composed mostly of other agents) form a system that plays the game with the external environment more capably than any agent? If you know of any active research programs with this flavor, I would love to hear from you (and you should feel welcome to email me).
  • I occasionally become absorbed by attempts to prove that PNPP \subsetneq NP implies or is implied by the fluctuation theorem. So far, it’s been a reliably humbling experience.
  • I’d like to better understand quantum decoherence. I never really outgrew a sophomoric concern with the Measurement Problem.
  • I would be keen to develop architectures for normalizing flows that allow exact or cheaply-approximated Fisher-Rao natural gradient descent. Such architectures would allow accelerated simulation of evolutionary dynamics as a consequence of my prior work relating natural gradient descent and evolutionary game theory {5}.


Machine Learning Multi-agent Systems, Reinforcement Learning, Nonstationary Environments.
Optimization Convex Optimization, Evolutionary Game Theory, Approximate Newton Methods.
Mathematics Information Theory, Linear Algebra, Vector Calculus, Variational Calculus.
Programming GNU/Linux, Python (incl. NumPy, SciPy, Jax, Taichi, Gym), C, Open-Source.


PhD in Computer Science and Engineering
Sept 2019 – Mar 2024
University of California, Santa Cruz
Santa Cruz, CA
  • ARCS Scholar
  • Dean’s Fellow
  • Regents Fellow
  • Advancement with Honors
  • Dissertation Year Fellow
BSc in Physics, College of Creative Studies
Sept 2011 – June 2015
University of California, Santa Barbara
Santa Barbara, CA
  • High Honors
  • Distinction in the Major
  • UC Education Abroad Program Scholarship
  • UCSB Education Abroad Program Scholarship


Paper Awards

Best Paper Runner-Up {6} RTML workshop at ICLR 2023
Highlighted Paper {6} RMTL Workshop at ICLR 2023
Spotlight Paper {3} NeurIPS 2021

Academic Honors

Advancement with Honors UC Santa Cruz 2021
High Honors (BSc in Physics) UC Santa Barbara 2015
Distinction in the Major (Physics) UC Santa Barbara 2015

Scholarships and Fellowships

Dissertation Year Fellowship (Winter) UC Santa Cruz 2024
Dissertation Year Fellowship (Fall) UC Santa Cruz 2023
Merit-Based Scholarship ARCS Foundation, Inc. (Northern California Chapter) 2022
Regents Fellowship UC Santa Cruz 2019
Dean's Fellowship UC Santa Cruz 2019
Education Abroad Scholarship UC (All Campuses) 2013
Education Abroad Scholarship UC Santa Barbara 2013
Undergraduate Research Internship UC Santa Barbara 2011

Work & Research Experience

Graduate Student Researcher
Sept 2019 – Mar 2024
Human-Centered Machine Learning
UC Santa Cruz

The Human-Centered Machine Learning Group at UC Santa Cruz researches the real-world, human consequences of deployed machine learning (ML) systems. My role in this group involves proposing original research questions, deriving theoretical results, designing numerical experiments, authoring computational simulations, writing papers and technical appendices, and presenting our research in multiple top conferences in talks and posters.

  • Described dynamics of systems of mutual learners using evolutionary game theory {3}.
  • Established adversarial bounds for fairness violation due to distribution shift {4}.
  • Discovered exact, least-squares correspondence between replicator dynamics and natural gradient descent {5}.
  • Adapted online reinforcement learning methods to systems of mutually interacting learners {6}.
  • Mapped machine learning with policy-induced distribution shift to a novel constrained optimization algorithm {7}.
  • Wrote multi-agent simulations in Python using JAX and Taichi for GPU acceleration.
Software Developer
Oct 2016 – Aug 2018
Breadware, Inc.
Reno, Nevada

As a startup company, Breadware, Inc. offered consulting and rapid prototyping for internet-of-things (IoT) products. My role at the company was to build proprietary tools to modularize device development and automate associated engineering tasks.

  • Mapped abstract hardware APIs to I2C bus protocols for modular embedded devices (Python, C).
  • Implemented web-based testing of user-logic for embedded devices in simulated environments (JavaScript).
  • Wrote scripts for automating electronic design tasks, such as PCB layout (Python).
Teaching Assistant and Residential Mentor
Summer 2015 | Summer 2016
The Summer Science Program
Socorro, NM | Boulder, CO

The Summer Science Program is a non-profit organization, run by its alumni (of which I am one), that, since 1959, has hosted advanced high school students at university campuses take accelerated college-level coursework and conduct research in observational astrophysics. I was a Teaching Assistant and Residential Mentor for the program in the summers of 2015 (at New Mexico Tech) and 2016 (at CU Boulder).

  • Mentored and supervised advanced high school students in observational astronomy.
  • Graded written and programming assignments in celestial mechanics, programming, and mathematics.
  • Wrote supplementary math and programming challenges and gave supplementary lectures.
Undergraduate Research Intern
Feb 2013 – Jun 2015
California NanoSystems Institute, UC Santa Barbara
Santa Barbara, CA

Under the California NanoSystems Institute, I was a paid undergraduate research intern for the Josephson Junction Superconducting Quantum Computing Group at UC Santa Barbara (now Google Quantum AI). My role included myriad tasks, in hardware, software, and calculation, to build one of the world’s first scalable universal quantum computers.

  • Fabricated a vibration-damped platform for a new liquid-helium refrigerator, superconducting cables and their brackets.
  • Designed a printed circuit board for a radio-frequency amplifier.
  • Wrote a Python package to automate phase-noise characterization of GHz voltage oscillators.
  • Calculated the effect of hardware signal-chain imperfections on quantum gate errors, enabling software compensation {2}.

    Undergraduate Research Intern
    Jan 2011 – Jul 2011
    UC Santa Barbara Center for Energy Efficient Materials
    Santa Barbara, CA

With a paid undergraduate research internship awarded by the UC Santa Barbara Center for Energy Efficient Materials, I worked for a group in the Chemistry Department on characterizing organic molecules for use in plastic solar cells.

  • Worked on organic electric device characterization to develop plastic solar cells.
  • Fit spectroscopic ellipsometry measurements to models for organic electron-donor molecules [1].
  • Conducted laser-induced photoluminescence decay measurements for same electron-donor molecules [1].
  • Performed atomic force microscopy measurements and statistical analyses for thin-layer morphology characterization.


{7} Fair Participation via Sequential Policies
Reilly Raab, Ross Boczar, Maryam Fazel, and Yang Liu
AAAI. (2024)

{6} Long-Term Fairness with Unknown Dynamics
Tongxin Yin[1], Reilly Raab[1], Mingyan Liu, and Yang Liu
ICLR RTML (2023) – Best Paper Runner-Up, Highlighted Paper Award
NeurIPS. (2023)

{5} Conjugate Natural Selection
Reilly Raab, Luca de Alfaro, and Yang Liu
arXiv. (2022)

{4} Fairness Transferability Subject to Bounded Distribution Shift
Yatong Chen[1], Reilly Raab[1], Jialu Wang, and Yang Liu
NeurIPS. (2022)

{3} Unintended Selection: Persistent Qualification Rate Disparities and Interventions
Reilly Raab and Yang liu
NeurIPS. (2021) – Spotlight Paper Award

{2} Single-Gate Error for Superconducting Qubits Imposed by Sideband Products of IQ Mixing
Reilly Raab
Undergraduate Thesis. (2015)

{1} Systematic Study of Exciton Diffusion Length in Organic Semiconductors by Six Experimental Methods
Jason D. A. Lin et. al.
Materials Horizons. (2014)


Equal contribution