Ke Liao

Yale University

prof_pic.jpg

I am a scientist and ML research engineer building physics-informed AI systems for scientific discovery. At Yale University, I work with Prof. Tianyu Zhu at the intersection of machine learning, quantum chemistry, and high-performance computing. My goal is to turn expensive, expert-driven scientific analysis into scalable and interpretable computational workflows that connect simulation with experiment.

What I build

  • AI that connects spectra to materials properties. I am developing AI4ARPES, a workflow for identifying magnetic phases directly from photoemission spectra. The work combines automated first-principles data generation, physics-informed neural-network classifiers, and domain-adaptation methods for bridging simulated and measured data.
  • GPU software for quantum many-body simulation. I lead the development of PyTC, a JAX-based framework for transcorrelated electronic-structure calculations. My recent work uses interpolative separable density fitting to compress the transformed Hamiltonian, bringing together numerical linear algebra, JIT-compiled accelerator kernels, and reproducible scientific-software design.
  • Learned representations of correlated electrons. I combine transcorrelated Hamiltonians with neural quantum states and quantum-embedding methods to build compact models of strongly correlated molecules and materials.

Across research and consulting, I translate scientific questions into computational systems: formulating learning problems, designing data-generation and validation pipelines, implementing models and numerical kernels, and evaluating accuracy and performance. As an external scientific consultant at ByteDance, I scoped and evaluated machine-learning research directions for quantum chemistry and advised on algorithmic choices and technical feasibility. My research experience at Yale, Caltech, LMU Munich, and the Max Planck Institute gives me the domain depth to work across ML, physics, and scientific software.

I am actively looking for Research Scientist, Applied Scientist, and ML Research Engineer roles in AI for science, scientific machine learning, and accelerated scientific computing. I am especially interested in teams building models and platforms that connect simulation, experiment, and scientific decision-making.