<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Minchan Jeong — Blog</title><description>Notes on scientific machine learning, JAX/XLA engineering, and operator-theoretic methods.</description><link>https://minchanjeong.github.io/</link><language>en-us</language><item><title>Algorithmic incompressibility, chaos, and the data-scaling ceiling in scientific machine learning</title><link>https://minchanjeong.github.io/blog/data-driven-sciml-ceiling/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/data-driven-sciml-ceiling/</guid><description>The big-data scaling recipe applies unevenly across scientific machine learning: it works for equilibrium tasks like AlphaFold but hits a structural ceiling at per-trajectory prediction of chaotic systems.</description><pubDate>Tue, 19 May 2026 00:00:00 GMT</pubDate></item><item><title>Dynamic batching, kernel switching, and the off-policy trap in RLVR</title><link>https://minchanjeong.github.io/blog/llm-inference-nondeterminism/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/llm-inference-nondeterminism/</guid><description>LLM inference nondeterminism is rarely a hardware artifact: dynamic batching switches kernels, the kernel switch reorders floating-point summations, and the resulting drift in logits silently turns RLVR training off-policy.</description><pubDate>Wed, 26 Nov 2025 00:00:00 GMT</pubDate></item><item><title>Koopman operator: adjoint, normality, SVD</title><link>https://minchanjeong.github.io/blog/koopman-adjoint-normality-svd/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/koopman-adjoint-normality-svd/</guid><description>What is the adjoint of the Koopman operator on a stochastic Markov process, when is the operator normal, and why does the singular value decomposition become the natural object the moment normality fails?</description><pubDate>Sat, 17 May 2025 00:00:00 GMT</pubDate></item><item><title>Spectral Theorem I: adjoint, normality, and the finite-dimensional case</title><link>https://minchanjeong.github.io/blog/adjoint-normality-spectral-theorem/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/adjoint-normality-spectral-theorem/</guid><description>Building the finite-dimensional spectral theorem from the basis-free adjoint, normality, and Gram–Schmidt induction on eigenvectors. Part I of a three-part series, with a separate Operator SVD capstone.</description><pubDate>Tue, 18 Feb 2025 00:00:00 GMT</pubDate></item><item><title>Spectral Theorem II: compact operators via the variational principle</title><link>https://minchanjeong.github.io/blog/compact-operator-spectral-theorem/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/compact-operator-spectral-theorem/</guid><description>The compact case between finite dimension and general bounded operators: the variational principle as the eigenvalue source, compactness itself as the discretizer that forces the eigenvalues to decay to zero. Part II of a three-part series, with a separate Operator SVD capstone.</description><pubDate>Tue, 18 Feb 2025 00:00:00 GMT</pubDate></item><item><title>Spectral Theorem III: general bounded operators via functional calculus</title><link>https://minchanjeong.github.io/blog/functional-analytic-spectral-theorem/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/functional-analytic-spectral-theorem/</guid><description>When a bounded self-adjoint operator has no eigenvectors at all, the eigenbasis is replaced by a projection-valued measure built from continuous functional calculus. Part III of a three-part series, with a separate Operator SVD capstone.</description><pubDate>Tue, 18 Feb 2025 00:00:00 GMT</pubDate></item><item><title>Operator SVD: from spectral theorem to polar form to singular values</title><link>https://minchanjeong.github.io/blog/operator-svd/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/operator-svd/</guid><description>The SVD as the spectral theorem applied to L*L via the polar form L = U|L|: why it is the optimal low-rank object, what shape it takes at each spectral-theorem rung, and where it stops existing.</description><pubDate>Tue, 18 Feb 2025 00:00:00 GMT</pubDate></item><item><title>Top-of-atmosphere incident solar radiation for operational GraphCast</title><link>https://minchanjeong.github.io/blog/tisr-operational-graphcast/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/tisr-operational-graphcast/</guid><description>The library default for TOA solar radiation was a different physical quantity than GraphCast&apos;s training data. How I diagnosed and reconstructed the correct one.</description><pubDate>Fri, 15 Mar 2024 00:00:00 GMT</pubDate></item><item><title>A staged single-node ETL pipeline for terabyte-scale weather data</title><link>https://minchanjeong.github.io/blog/staged-single-node-etl/</link><guid isPermaLink="true">https://minchanjeong.github.io/blog/staged-single-node-etl/</guid><description>Materialized intermediates, fit-once transforms, executor per stage, quarantined dependencies: four design decisions in a single-node ETL for terabyte-scale weather data, with the alternatives I rejected.</description><pubDate>Fri, 01 Dec 2023 00:00:00 GMT</pubDate></item></channel></rss>