Hello, I'm Tae-Geun Kim 👋

🙋‍♂️ Introduce myself

👨‍🏫 Postdoctoral Researcher in Physics

• Institute of Modern Physics, Fudan University & RIKEN iTHEMS
• Curriculum Vitae · Google Scholar · ORCID
• Blog

🔬 Research Areas

• AI for Science — neural operators, physics-informed deep learning, Hamiltonian learning
• Dark matter physics — axion-like particles, primordial black holes, detectability studies
• Scientific & High Performance Computing — numerical algorithms, Rust-based tools, parallel computing

💼 Portfolio

Peroxide

• Comprehensive Rust numeric library for linear algebra, numerical analysis, statistics, and machine learning
• Supports automatic differentiation, special functions, DataFrame, and BLAS/LAPACK integration
• User-friendly syntax inspired by R, NumPy, and MATLAB

HyperbolicLR

• Novel learning rate schedulers addressing the learning curve decoupling problem in deep learning
• Epoch-insensitive design enables stable training across varying durations without retuning
• Evaluated on image classification, time series prediction, and operator learning tasks with PyTorch

arXiv Explorer

• Personal research assistant for arXiv — discover, organize, and annotate papers from the terminal
• TF-IDF content similarity + category/keyword/recency scoring for personalized recommendations
• Full TUI, AI summaries (Gemini/Claude/OpenAI/Ollama), reading lists, and export to Markdown/JSON/CSV

Puruspe

• Pure Rust special functions library (gamma, beta, error functions) with zero dependencies
• Lightweight implementation based on "Numerical Recipes" algorithms

Forger

• Reinforcement Learning library in Rust with modular agents, environments, and policies
• Includes Epsilon Greedy, Value Iteration, and Q-Learning implementations

More projects

PyTorch Template

• Flexible PyTorch experiment template with YAML-based configuration
• Supports multiple seeds, device selection, and LR scheduling for reproducible ML research

Radient

• Rust automatic differentiation library using computational graphs
• Forward/backward propagation with cached and non-cached gradient options

DeeLeMa

• Deep learning network for mass and momentum estimation in high-energy collider events
• Robust mass peak recovery under combinatoric uncertainties and detector smearing

📚 Publications

• Yongsoo Jho, Tae-Geun Kim, Jong-Chul Park, Seong Chan Park and Yeji Park, Primordial Black Holes as a Factory of Axions: Extragalactic Photons from Axions, Prog. Theor. Exp. Phys. ptag011, arXiv:2212.11977 (2026)

• Taehyeun Kim, Tae-Geun Kim, Anouk Girard, Ilya Kolmanovsky, Learning Hamiltonian Dynamics with Bayesian Data Assimilation, arXiv:2501.18808 (2025)

• Tae-Geun Kim, Seong Chan Park, Neural Hamilton: Can A.I. Understand Hamiltonian Mechanics?, arXiv:2410.20951 (2024)

• Tae-Geun Kim, HyperbolicLR: Epoch insensitive learning rate scheduler, arXiv:2407.15200 (2024)

• Chang Min Hyun, Tae-Geun Kim, and Kyounghun Lee, Unsupervised sequence-to-sequence learning for automatic signal quality assessment in multi-channel electrical impedance-based hemodynamic monitoring, CMPB 108079, arXiv:2305.09368 (2023)

• Kayoung Ban, Dong Woo Kang, Tae-Geun Kim, Seong Chan Park and Yeji Park, DeeLeMa : Missing information search with Deep Learning for Mass estimation, Phys. Rev. Research 5, 043186, arXiv:2212.12836 (2022)

🔖 Skills

• Primary Languages : Rust, Python, C++, Julia
• Frameworks & Libraries
  • Numerical Computing: peroxide, numpy, scipy, pandas/polars, BLAS/LAPACK, eigen, mathematica
  • Machine Learning: PyTorch, JAX/Equinox/Optax, W&B, Optuna, Candle, TensorFlow, Scikit-Learn
  • Visualization: matplotlib, vegas, ggplot2, plotly
  • High Energy Physics: BlackHawk, GALPROP, MadGraph, ROOT
  • Quantum Computing: PennyLane, Qiskit, Cirq, RustQIP
  • Web: Django, Vue, Firebase, Hugo, Zola, Elm

Github contributions