← Back to People

Dr. Junwen Yin

Computational Scientist(Senior)
STFC SCD, UKRI | ETSF Member
Dr. Junwen Yin

Biography

Junwen is the lead developer of Goldilocks, which uses high-throughput DFT, machine learning and LLM-based AI agents to choose converged, cost-efficient settings for DFT calculations, so that less compute and electricity is wasted on national supercomputers such as ARCHER2.

Goldilocks turns a crystal structure into recommended calculation settings, records the basis for each recommendation, and generates the input files and submission script needed to run it. It draws on machine-learning models and datasets published on PSDI, including converged k-point meshes for more than 17,000 structures. The tool runs as a command-line program, a browser application, an HTTP API and an MCP server for AI assistants; an LLM-based agent, in development, orchestrates scientific workflows and explains its recommendations in plain language. The k-point work is published in Digital Discovery (2026).

Alongside Goldilocks, Junwen works on first-principles methods and their applications: electron–phonon coupling and phonon anharmonicity from DFT and DFPT, including carrier mobility in halide perovskites and the phonon-limited resistivity of metals; nonadiabatic coupling and excited-state dynamics in CP2K; and ferroelectric hafnia.

Highlighted Work

💻 Software goldilocks.ac.uk

Recommends the settings for a DFT calculation from a crystal structure, records the basis for each recommendation, and generates the input files and submission script, so calculations are neither under-converged nor wasteful. Built as a data-driven pipeline, with funding from the Ada Lovelace Centre and EPSRC (EP/Z530657/1), to cut compute and electricity waste on national supercomputers such as ARCHER2.

Research Interests & Focus

  • High-throughput DFT and automated convergence of calculation parameters (Goldilocks)
  • Machine learning, LLMs and AI agents for electronic-structure workflows
  • Research software for automated DFT: command-line tools, web applications, APIs and MCP servers
  • Electron–phonon coupling, phonon anharmonicity and carrier transport from DFT and DFPT
  • Halide perovskites and ferroelectric hafnia
  • Nonadiabatic coupling and excited-state dynamics in CP2K
  • Catalytic, electrochemical and photoactive interfaces
Goldilocks High-Throughput Calculations Machine Learning LLMs AI Agents Research Software DFT DFPT Electron–Phonon Coupling Nonadiabatic Dynamics Perovskites Catalytic Interfaces CP2K Framework

Appointments & Roles

Recent Publications

An Efficient On-the-Fly Nonadiabatic Coupling Framework Integrated into CP2K
Xiaoke He, Hsiao-Yi Tsai, Ziwei Chai, Junwen Yin, Zhongfei Xu, Li-Min Liu The Journal of Physical Chemistry Letters • 2026
Automatic generation of input files with optimised k-point meshes for Quantum ESPRESSO self-consistent field single-point total energy calculations
Elena Patyukova, Junwen Yin, Susmita Basak, Samuel Pinilla, Alin M. Elena, Gilberto Teobaldi Digital Discovery • 2026 DOI: 10.1039/d5dd00565e
Data-driven Design of Metal-Organic Frameworks with Tunable Negative Thermal Expansion
Prathami Divakar Kamath, Francesco Tavani, Alin Marin Elena, Théo Jaffrelot Inizan, Yen-hsu Lin, Jian Yin, Wenqian Xu, Omar M. Yaghi, Kristin A. Persson arXiv preprint arXiv:2607.18594 • 2026 DOI: 10.48550/arXiv.2607.18594
Identify the switching of VO 3 and VO 4 and its effect on ferroelectric hafnia
Sida Li, Zihe Wang, Tingxiao Xie, Chuang Xue, Yufeng Xue, Qian Yin, Tongcai Yue, Junwen Yin, Tengfei Cao, Gilberto Teobaldi, Qi Hu, Li-Min Liu Physical Review B • 2026 DOI: 10.1103/2j8c-3z3h
Opposite impact of thermal expansion and phonon anharmonicity on the phonon-limited resistivity of elemental metals from first principles
Ao Wang, Junwen Yin, Félix Antoine Goudreault, Michel Côté, Olle Hellman, Samuel Poncé Physical Review B • 2026 DOI: 10.1103/yk57-bt6t
View All Publications by Dr. Junwen Yin (16) →