Mikel Landajuela
I lead AI-for-science and scientific-computing teams at Lawrence Livermore National Laboratory in the Computational Engineering Directorate, building machine learning systems for antibody design, interpretable modeling, and high-performance simulation. My work spans research leadership and hands-on development across machine learning, reinforcement learning, optimization, and HPC, with work published in Nature, NeurIPS, ICML, and AAAI. I earned my Ph.D. from Universit茅 Pierre et Marie Curie and Inria in Paris, and received the SMAI-GAMNI award 2017 from the French Society of Industrial and Applied Mathematics.
Expertise: machine learning 路 reinforcement learning 路 optimization 路 scientific computing 路 mathematical modeling
Projects & Focus Areas
AI for Science & Industry : Developing data-driven and physics-based algorithms on GPU/HPC clusters. Currently building deep-learning pipelines for protein design and medical countermeasures.
Interpretable & Responsible AI : Crafting symbolic-regression methods and decision-tree controllers to ensure transparent, auditable ML workflows suitable for regulated environments.
High-Performance Scientific Computing : Designing scalable PDE solvers for fluid-structure interaction and cardiovascular mechanics, optimizing C++/Python code across on-prem and cloud HPC environments.
I welcome research collaborations and industry partnerships. 鉁夛笍 landajuelala1 at llnl dot gov 路 馃敆 Google Scholar
Selected Highlights
- 2025-09-23: Keynote speaker at AI for SCIENCE 2025, Ljubljana, Slovenia.
- 2025-03-28: Preemptive optimization of a clinical antibody for broad neutralization of SARS-CoV-2 variants and robustness against viral escape published in Science Advances.
- 2024-05-08: Computationally restoring the potency of a clinical antibody against Omicron published in Nature.
- 2022-09-14: A Unified Framework for Deep Symbolic Regression accepted at NeurIPS 2022. (openreview)