Jekyll2026-08-05T07:20:39+00:00https://pyiron.org/feed.xmlpyironAn interactive development environment (IDE) for computational materials science.{"name"=>nil, "email"=>nil, "twitter"=>nil}Coupling High-Throughput Density Functional Theory, Automated Experimentation, and Adaptive Experimental Design To Achieve Selective Rare-Earth Element Separations2026-07-21T00:00:00+00:002026-07-21T00:00:00+00:00https://pyiron.org/publications/2026/07/21/high-throughput-separationLogan J. AugustineExecutorlib Tutorial at Los Alamos National Laboratory2026-05-27T00:00:00+00:002026-05-27T00:00:00+00:00https://pyiron.org/news/workshop/2026/05/27/lanl-executorlib-tutorialWith the increasing popularity of the executorlib library to up-scale python functions for high performance computing (HPC) the pyiron developers organized an executorlib tutorial at Los Alamos National Laboratory. The tutorial covered an introduction to job schedulers starting with the Simple Linux Utility for Resource Management (SLURM) and the hierarchical job scheduler flux framework. Followed by an extensive introduction to executorlib with a focus on up-scaling Python functions from local workstations, to local high performance computers (HPC) and finally the latest generation of Exascale computers.

The content of the tutorial is available on Github.

]]>{"name"=>nil, "email"=>nil, "twitter"=>nil}Hierarchical Multi-agent Large Language Model Reasoning for Autonomous Heterogeneous Catalyst Discovery2026-05-25T00:00:00+00:002026-05-25T00:00:00+00:00https://pyiron.org/publications/2026/05/25/llm-catalystSamuel RothfarbToward Knowledge-Based Workflows: A Semantic Approach to Atomistic Simulations for Mechanical and Thermodynamic Properties2026-05-12T00:00:00+00:002026-05-12T00:00:00+00:00https://pyiron.org/publications/2026/05/12/knowledge-based-workflowsAbril Azócar GuzmánDFT and MLIP study of solute segregation to coherent and semi-coherent α-Fe/Fe3C interfaces2026-04-03T00:00:00+00:002026-04-03T00:00:00+00:00https://pyiron.org/publications/2026/04/03/semi-coherent-iron-interfacesAmin Reiners-SakicTutorial at German Physics Conference2026-03-08T00:00:00+00:002026-03-08T00:00:00+00:00https://pyiron.org/news/workshop/2026/03/08/dpg-tutorialIn the tutorial titled Linking large language models with digital workflows for materials science simulations the pyiron developers presented the recent advances of the pyiron workflow framework, with a specific focous on the integration of large language models (LLM) to accelerate the workflow development.

The tutorial website summarizes the topics covered in the tutorial and the corresponding content is available in the Github repository.

In comparison to generating Python code directly with an LLM the advantage of using a workflow framework is the restriction of the flexibility of to focus on the generation of nodes and edges which reduces the risk of hallucination.

]]>{"name"=>nil, "email"=>nil, "twitter"=>nil}Augmenting Large Language Models for Automated Discovery of F-Element Extractants2026-01-28T00:00:00+00:002026-01-28T00:00:00+00:00https://pyiron.org/publications/2026/01/28/llm-discovery-of-f-elementsBaosen ZhangMachine-learning interatomic potential for BaTiO3: phase transitions, domain walls, and grain boundaries2025-12-18T00:00:00+00:002025-12-18T00:00:00+00:00https://pyiron.org/publications/2025/12/18/BaTiO3Amit SehrawatDesign of active sites for amine-functionalized direct air capture materials using integrated high-throughput calculations and machine learning2025-12-08T00:00:00+00:002025-12-08T00:00:00+00:00https://pyiron.org/publications/2025/12/08/direct-air-capture-materialsMegan C. DavisPython Workflow Definition2025-11-18T00:00:00+00:002025-11-18T00:00:00+00:00https://pyiron.org/news/releases/2025/11/18/python-workflow-definitionIn collaboration with the developers of aiida and jobflow the pyiron developer team released the Python Workflow Definition an interoperability format for Python based workflows. In contrast to previous standardization efforts like the Common Workflow Language the Python Workflow Definition is based on sharing Python objects between consecutive workflows steps, rather than files. This provides a performance advantage in particular for high-throughput workflows as they are commonly used in computational materials science.

Read the publication about the Python Workflow Definition on Digital Discovery.

]]>{"name"=>nil, "email"=>nil, "twitter"=>nil}