About me
Hi, I’m Dr. Ziqi Guo, a Machine Learning Researcher at Apple working on battery modeling and simulation. My work sits at the intersection of AI for Science, multiscale physics modeling, and high-performance computing: I build physics-informed machine learning models and LLM-agent workflows that turn slow, expensive simulations and manual engineering processes into fast, reliable tools for energy, thermal management, and sustainability.
I received my Ph.D. in Mechanical Engineering from Purdue University (Aug. 2025, GPA 3.97/4.00), co-advised by Prof. Xiulin Ruan and Prof. Guang Lin. Before that, I earned my B.E. with Honors in Energy & Power Engineering from Huazhong University of Science and Technology (rank 1/221, GPA 3.99/4.00).
Current work
Machine Learning Researcher, Battery Modeling & Simulation, Apple (Jun. 2025 – present)
- Developed a physics-informed ML framework that combines physics priors with a fine-tuned transformer-based tabular foundation model to predict the performance of new battery cell designs across form factors, with 30% lower error than the empirical scaling rules used previously.
- Applied the framework to a next-generation cell chemistry ahead of hardware builds (predictions within 15% of measurements), and combined it with physics-based open-circuit-voltage, impedance, and aging models calibrated to cell test data to evaluate control algorithms and design feasibility across three product lines.
- Built an end-to-end LLM agent pipeline that extracts and validates inputs from design documents and cell test data, runs the ML framework to generate full cell models, and drafts reports that flag risks, cutting turnaround from one week to one day.
- Research intern (May – Aug. 2024): developed numerical solvers for an on-device physics-based lithium-ion battery model (16× faster simulation) and a time-series model for on-device user-behavior forecasting (50% higher accuracy than the existing model).
Research background
During my Ph.D. I developed ML and HPC methods that accelerate first-principles simulation of how heat and light move through materials.
- Fast phonon transport. Built an ML model that predicts phonon scattering rates and lattice thermal conductivity at first-principles accuracy with up to 100× speedup, and a statistical-sampling method that makes converged calculations tractable. Then built FourPhonon_GPU, a CPU–GPU heterogeneous scheme (MPI/OpenMP + OpenACC) that gives up to 26× overall speedup (400× for the four-phonon scattering step). All are released in the open-source FourPhonon package and were featured twice by Purdue Engineering.
- Atomistic foundation models. Fine-tuned a pretrained universal ML force field for Al/Si interfaces, reaching ab initio accuracy (RMSE < 0.8 meV/atom) with ~100× faster molecular dynamics.
- Physics-informed neural networks. Developed PINNs for forward and inverse heat transfer, recovering thermal diffusivity from sparse temperature data with < 5% relative error.
- Radiative cooling and optics. Co-developed the open-source FOS Monte Carlo photon-transport solver, built an RNN surrogate that is 700× faster than Monte Carlo for multi-layer media (< 1% error), and designed a generative inverse-design pipeline that nearly doubles the visible reflectance of colored radiative cooling paints (40% → 76% at matched color).
- Materials physics. First-principles studies of hexagonal boron nitride (thermal conductivity, Raman/IR linewidths) and its use as a radiative cooling pigment, and of β-Ga2O3 thermal transport.
In total I have published 20+ papers (see Publications or my Google Scholar profile), with 560+ citations and an h-index of 12.
Skills
- Machine learning: physics-informed ML, ML force fields, graph and recurrent neural networks, surrogate modeling, inverse design, time-series forecasting, foundation-model fine-tuning (PyTorch, TensorFlow, scikit-learn)
- LLM agents: information extraction, workflow automation (LangChain, LangGraph, Claude Agent SDK)
- Physics modeling: battery equivalent-circuit and aging models, DFT, molecular dynamics, finite element analysis
- Programming & HPC: Python, C/C++, Fortran, MATLAB, SQL; MPI, OpenMP, OpenACC, CUDA, SLURM
Selected honors
- Dimitris N. Chorafas Foundation Award (2025): about 20 doctoral recipients worldwide
- Graduate Student Gold Award, Materials Research Society (2025): MRS’s highest graduate student honor
- Magoon Research Excellence Award, Purdue University (2025): four recipients per year
- Warren and Judy Stevenson Scholarship for outstanding research in applied optics (2024)
- Best / Outstanding Presentation Awards at ASME conferences (3×, 2023–2024)
- National Scholarship (3×) and Exemplary Merit Student Award, HUST
More in News.
Recent news
Selected publications
Ziqi Guo, Dudong Feng, Daniel Carne, Guang Lin, and Xiulin Ruan. “Machine Learning Inverse Design Reveals a Double Narrow-Band Absorption Approach for Effective Colored Radiative Cooling Paints.” Nano Letters 26 (2026): 8359-8367. Link, Paper
Ziqi Guo, David McLoughlin, Xiulin Ruan, and Guang Lin. “FourPhonon_GPU, a CPU-GPU heterogeneous computing scheme for phonon scattering and thermal conductivity calculations.” Cell Reports Physical Science 7 (2026): 103491. Link, Paper
Ziqi Guo, Daniel Carne, Krutarth Khot, Dudong Feng, Guang Lin, and Xiulin Ruan. “A Review of Artificial Intelligence-Driven Approaches for Nanoscale Heat Conduction and Radiation.” Journal of Computing and Information Science in Engineering 25, no. 12 (2025): 120805. Link, Paper
Ziqi Guo, Ioanna Katsamba, Daniel Carne, Dudong Feng, Kellan Moss, Emily Barber, Ziqi Fang, Andrea Felicelli, and Xiulin Ruan. “Electronic and phononic characteristics of high-performance radiative cooling pigments h-BN: A comparative study to BaSO4.” Materials Today Physics (2025): 101721. Link, Paper
Ziqi Guo, Peter Sokalski, Zherui Han, Yanhua Cheng, Li Shi, Takashi Taniguchi, Kenji Watanabe, and Xiulin Ruan. First-principles prediction of Raman and infrared linewidths of hexagonal boron nitride. Applied Physics Letters, Dec, 2024. Link, Paper
Ziqi Guo, Zherui Han, Abdulaziz Alkandari, Krutarth Khot and Xiulin Ruan. First-principles prediction of thermal conductivity of bulk hexagonal boron nitride. Applied Physics Letters, Apr, 2024. Link, Paper
Ziqi Guo, Zherui Han, Dudong Feng, Guang Lin, and Xiulin Ruan. Sampling-accelerated first-principles prediction of phonon scattering rates for converged thermal conductivity and radiative properties (2023). npj Computational Materials, Feb, 2024. Link, Paper
Ziqi Guo, Prabudhya Roy Chowdhury, Zherui Han, Yixuan Sun, Dudong Feng, Guang Lin, and Xiulin Ruan. Fast and Accurate Machine Learning Prediction of Phonon Scattering Rates and Lattice Thermal Conductivity. npj Computational Materials, Jun, 2023. Link, Paper
Sijie Li*, Ziqi Guo*, Jacob B Ioffe, Yunfei Hu, Yi Zhen, Xin Zhou. Autism_genepheno: Text mining of gene-phenotype associations reveals new phenotypic profiles of autism-associated genes. Scientific Reports, Jul. 2021. Link, Paper