PeopleofSuperintelligence

Researcher

Justus C. Will

Also known as Justus Will

Justus C. Will is a PhD student in machine learning at the University of California, Irvine, working on climate-model emulation, neural data compression and faster decoding for language models. After degrees in mathematics and computer science at TU Kaiserslautern, he worked there as a research assistant in Marius Kloft's Machine Learning Group on tensor completion. He began his PhD at UC Irvine in 2023, advised by Stephan Mandt, and was an author of ClimSim, a dataset of 5.7 billion input-output pairs for training machine-learning climate emulators that won an Outstanding Paper Award at NeurIPS 2023. In 2025, his paper with Yibo Yang and Mandt on progressive image compression with universally quantized diffusion models was selected as an oral at ICLR, and in 2026 he presented Parallel Token Prediction for language models at the same conference.

Timeline

  1. 2017–2022

    Degrees at TU Kaiserslautern

    He studies mathematics and computer science at TU Kaiserslautern, earning two bachelor's degrees and then an M.Sc. in mathematics in 2022, with a semester abroad at Lund University in Sweden.

  2. 2019–2022

    Tensor completion in Kaiserslautern

    From October 2019 to May 2020, he is a research assistant at the German Research Center for Artificial Intelligence (DFKI), where he develops an evolutionary algorithm that optimizes the topology and hyperparameters of convolutional networks. From October 2020 to December 2022, he is a research assistant in Marius Kloft's Machine Learning Group at TU Kaiserslautern, where he develops a tensor completion framework for sparse tabular data.

  3. 2023

    A PhD at UC Irvine

    In 2023, he begins a Ph.D. in computer science at the University of California, Irvine, advised by Stephan Mandt.

  4. 2023

    ClimSim wins a NeurIPS award

    He is an author of ClimSim, a dataset of 5.7 billion pairs of input and output vectors for training machine-learning emulators of high-resolution climate physics, developed by a consortium of climate scientists and ML researchers. The paper receives an Outstanding Paper Award in the Datasets and Benchmarks Track at NeurIPS 2023.

  5. 2025

    ClimSim-Online in JMLR

    In 2025, the Journal of Machine Learning Research publishes ClimSim-Online, on which he is an author. It extends ClimSim with an end-to-end workflow and a containerized pipeline for coupling ML models into operational climate simulators for hybrid testing.

  6. 2025

    Progressive compression at ICLR

    In 2025, he and Yibo Yang, with Stephan Mandt, present Progressive Compression with Universally Quantized Diffusion Models, selected as an oral at ICLR. The paper proposes a diffusion model with uniform noise in the forward process, whose negative ELBO corresponds to the end-to-end compression cost under universal quantization.

  7. 2026

    Parallel Token Prediction at ICLR

    In 2026, he coauthors Parallel Token Prediction for Language Models with Felix Draxler, Stephan Mandt and others, presented at ICLR in April 2026. The method predicts several tokens in a single model call by moving the source of randomness from post-hoc sampling to random input variables.

  8. As of October 2026

    PhD student at UC Irvine

    He is a PhD student at the University of California, Irvine, advised by Stephan Mandt, with research on deep generative models, neural data compression and machine learning for climate science.

Career

  1. 2020–2022University of KaiserslauternResearch Assistant
  2. 2019–2020DFKIResearch Assistant

Education

  1. From 2023University of California, IrvinePhD
  2. 2022University of KaiserslauternBachelor's degrees in Mathematics and Computer Science; M.Sc. in Mathematics

From the biography.

  1. Nonprofit roles are 2.3 times as common in AI safetyIncludes DFKI. We estimate where people focused on safety, alignment and evaluation work. Nonprofit roles appear for 25.1% of safety people with known affiliations, compared with 11.0% in other areas.
  2. The top 10 organizations employed 22% of people shaping AI in 2025: 51% of company researchers and 21% of university researchersIncludes DFKI. We estimate the share of people shaping AI who worked at the top 10 organizations each year from 1955 to 2025, comparing concentration across companies, universities, and institutes and labs.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes University of Kaiserslautern. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  4. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes University of Kaiserslautern. 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.

Colleagues

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

Sources and credits

Photo: We have no published photo of Justus C. Will yet. Send us a free-licensed photo.

  1. Justus Will justuswill.com. Checked 21 October 2026.
  2. Justus Will - Machine Learning Group, TU Kaiserslautern ml.cs.uni-kl.de. Checked 21 October 2026.
  3. #NeurIPS2023 outstanding papers - AIhub aihub.org. Checked 21 October 2026.
  4. ClimSim: A large multi-scale dataset for hybrid physics-ML climate emulation proceedings.neurips.cc. Checked 21 October 2026.
  5. ClimSim-Online: A Large Multi-scale Dataset and Framework for Hybrid ML-physics Climate Emulation arxiv.org. Checked 21 October 2026.
  6. Progressive Compression with Universally Quantized Diffusion Models arxiv.org. Checked 21 October 2026.
  7. ICLR 2025 Oral: Progressive Compression with Universally Quantized Diffusion Models iclr.cc. Checked 21 October 2026.
  8. Parallel Token Prediction for Language Models proceedings.iclr.cc. Checked 21 October 2026.
  9. ICLR Poster Parallel Token Prediction for Language Models iclr.cc. Checked 21 October 2026.