PeopleofSuperintelligence

Researcher

Joseph Kleinhenz

Joseph Kleinhenz is a researcher whose work spans quantum many-body simulation and generative artificial intelligence for molecular and protein design. In computational physics, he coauthored papers on inchworm quantum Monte Carlo for nonequilibrium impurities in February 2017, core ALPS simulation libraries in November 2018, and nonequilibrium metal-insulator transitions in the Hubbard model in April 2020. Turning to generative biology, he coauthored Protein Discovery with Discrete Walk-Jump Sampling for antibody proteins in June 2023, which won an ICLR 2024 Outstanding Paper Award, followed by Concept Bottleneck Protein Language Models in November 2024 and the JAMUN and FuncBind molecular generation frameworks in 2025. He is affiliated with Prescient Design, Genentech Roche, where he coauthored Diffeomorphic Optimization in July 2026 to map gradient-based optimization from a learned data manifold to a simpler base space.

Now at
Roche

Timeline

  1. February 2017

    Inchworm quantum Monte Carlo

    In February 2017, he coauthors Currents and Green's functions of impurities out of equilibrium, introducing a method to compute Green's functions, spectral functions, and currents for inchworm Monte Carlo.

  2. November 2018

    Core ALPS simulation libraries

    In November 2018, he coauthors Updated Core Libraries of the ALPS Project, presenting new Monte Carlo libraries and new Green's function libraries for physics simulation.

  3. November 2019

    Kondo voltage splitting dynamics

    His November 2019 paper Dynamics of Kondo voltage splitting after a quantum quench models a two-stage process in which a single central Kondo peak forms and then splits.

  4. April 2020

    A Hubbard-model phase transition

    In April 2020, he coauthors Dynamic Control of Nonequilibrium Metal-Insulator Transitions, which demonstrates a first-order metal-insulator phase transition in the repulsive, fully frustrated, single-band Hubbard model as a function of its coupling to a fermion bath.

  5. June 2023

    Discrete Walk-Jump Sampling for antibodies

    In June 2023, he coauthors Protein Discovery with Discrete Walk-Jump Sampling, combining an energy-based model with a score-based model at a single noise level for generative modeling of antibody proteins. In June 2023, he also coauthors 3D molecule generation by denoising voxel grids, a score-based method that represents 3D molecules as atomic densities on regular grids.

  6. 2024

    ICLR award and quasi Monte Carlo

    In 2024, his paper Protein Discovery with Discrete Walk-Jump Sampling wins an ICLR 2024 Outstanding Paper Award. In September 2024, while affiliated with Lawrence Berkeley National Laboratory, he coauthors Inchworm quasi Monte Carlo for quantum impurities, showing that imaginary-time integration with quasi Monte Carlo improves convergence. In November 2024, he coauthors Concept Bottleneck Language Models For protein design, introducing generative masked protein language models in which each neuron in a layer corresponds to an interpretable concept.

  7. 2025

    JAMUN and FuncBind at Genentech

    In March 2025, he is affiliated with Genentech and coauthors a JAMUN presentation at the APS Global Physics Summit. His 2025 NeurIPS paper JAMUN uses walk-jump sampling and score-based learning to generate conformational ensembles for small peptides an order of magnitude faster than traditional molecular dynamics. In November 2025, he coauthors Unified all-atom molecule generation with neural fields, which introduces FuncBind to represent molecules as continuous atomic densities.

  8. July 2026

    Diffeomorphic Optimization at Prescient Design

    In July 2026, he coauthors Diffeomorphic Optimization while affiliated with Prescient Design, Genentech Roche. The method uses diffusion and flow models to provide a map from a learned data manifold to a simpler base space for gradient descent.

Career

  1. Since 2025Roche
  2. 2024Office of Science

From the biography.

  1. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Roche. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  2. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Roche. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  3. Google had the most people shaping AI every year from 2014 to 2025Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  4. The United States employs an estimated 48% of people shaping AIWe estimate where the people shaping AI work today, how universities, companies and research institutes split across countries, and how country shares have changed since the 1960s.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Joseph Kleinhenz yet. Send us a free-licensed photo.

  1. Currents and Green's functions of impurities out of equilibrium: Results from inchworm quantum Monte Carlo journals.aps.org. Checked 22 October 2026.
  2. [1811.08331] Updated Core Libraries of the ALPS Project arxiv.org. Checked 22 October 2026.
  3. Dynamics of Kondo voltage splitting after a quantum quench | Phys. Rev. B journals.aps.org. Checked 22 October 2026.
  4. Dynamic Control of Nonequilibrium Metal-Insulator Transitions arxiv.org. Checked 22 October 2026.
  5. Protein Discovery with Discrete Walk-Jump Sampling arxiv.org. Checked 22 October 2026.
  6. [2306.07473] 3D molecule generation by denoising voxel grids arxiv.org. Checked 22 October 2026.
  7. ICLR 2024 Outstanding Paper Awards blog.iclr.cc. Checked 22 October 2026.
  8. Inchworm quasi Monte Carlo for quantum impurities journals.aps.org. Checked 22 October 2026.
  9. [2411.06090] Concept Bottleneck Language Models For protein design arxiv.org. Checked 22 October 2026.
  10. Machine Learning for Complex Systems - APS Global Physics Summit 2025 meetings-archive.aps.org. Checked 22 October 2026.
  11. JAMUN: Bridging Smoothed Molecular Dynamics and Score-Based Learning for Conformational Ensemble Generation papers.neurips.cc. Checked 22 October 2026.
  12. [2511.15906] Unified all-atom molecule generation with neural fields arxiv.org. Checked 22 October 2026.
  13. Diffeomorphic Optimization arxiv.org. Checked 22 October 2026.