{
  "schema_version": "1.0",
  "url": "https://cerinamroth.com/",
  "title": "Cerin Amroth — ML, JWST & Security Research",
  "description": "Independent research in 4-bit Mixture-of-Experts models, JWST little red dots and software security. Explore open-source tools, telescope images and evidence.",
  "date_modified": "2026-10-24",
  "representations": {
    "json": "https://cerinamroth.com/data/pages/home.json",
    "markdown": "https://cerinamroth.com/data/pages/home.md"
  },
  "headings": [
    {
      "level": 1,
      "id": null,
      "text": "Cerin Amroth — independent research: ML-systems efficiency, JWST observational work, and software security"
    },
    {
      "level": 2,
      "id": null,
      "text": "ML systems — large models, smaller machines"
    },
    {
      "level": 2,
      "id": null,
      "text": "Start with the problem"
    },
    {
      "level": 2,
      "id": null,
      "text": "Astronomy — explore real JWST observations"
    },
    {
      "level": 2,
      "id": null,
      "text": "Security — find the flaw, help get it fixed"
    }
  ],
  "links": [
    {
      "text": "Home",
      "url": "https://cerinamroth.com/"
    },
    {
      "text": "ML",
      "url": "https://cerinamroth.com/ml/"
    },
    {
      "text": "JWST",
      "url": "https://cerinamroth.com/jwst/"
    },
    {
      "text": "Security",
      "url": "https://cerinamroth.com/policy/"
    },
    {
      "text": "Contact",
      "url": "https://cerinamroth.com/#contact"
    },
    {
      "text": "Loggetta Check your hardware, plan a training run, and keep the results. One install includes the runtime and kernels. v0.7.0 · Plan and run training · pip install loggetta →",
      "url": "https://cerinamroth.com/ml/"
    },
    {
      "text": "experts4bit-qlora Fine-tune and serve large MoE models on one GPU. Move weights to RAM or SSD when GPU memory runs out. v0.58.1 · Get started →",
      "url": "https://cerinamroth.com/ml/experts4bit-qlora/"
    },
    {
      "text": "grouped-nf4-gemm Compute directly on packed 4-bit weights. GPU kernels, memory tools, and tests you can reproduce. v0.51.0 · Get started →",
      "url": "https://cerinamroth.com/ml/grouped-nf4-gemm/"
    },
    {
      "text": "OpenVINO Upstream fixes keeping Intel-GPU LLM inference off the CPU-fallback path — three merged.",
      "url": "https://cerinamroth.com/ml/openvino/"
    },
    {
      "text": "Latest ML research · 8 October 2026",
      "url": "https://cerinamroth.com/research/#update-2026-10-08"
    },
    {
      "text": "GitHub: Loggetta",
      "url": "https://github.com/pjordanandrsn/loggetta"
    },
    {
      "text": "PyPI",
      "url": "https://pypi.org/project/loggetta/"
    },
    {
      "text": "models",
      "url": "https://pypi.org/project/loggetta/#models"
    },
    {
      "text": "measured speed and architecture",
      "url": "https://cerinamroth.com/ml/"
    },
    {
      "text": "lower-layer solution guides",
      "url": "https://cerinamroth.com/ml/solutions/"
    },
    {
      "text": "Loggetta",
      "url": "https://cerinamroth.com/ml/"
    },
    {
      "text": "experts4bit-qlora internals",
      "url": "https://cerinamroth.com/ml/solutions/bitsandbytes-moe-load-in-4bit-still-ooms/"
    },
    {
      "text": "experts4bit-qlora",
      "url": "https://cerinamroth.com/ml/solutions/qlora-fused-moe-experts/"
    },
    {
      "text": "e4b residency paths",
      "url": "https://cerinamroth.com/ml/solutions/run-moe-larger-than-vram/"
    },
    {
      "text": "grouped-nf4-gemm",
      "url": "https://cerinamroth.com/ml/solutions/nf4-grouped-gemm-without-bf16-materialization/"
    },
    {
      "text": "grouped-nf4-gemm",
      "url": "https://cerinamroth.com/ml/solutions/native-mxfp4-moe-inference/"
    },
    {
      "text": "Receipts-driven engineering",
      "url": "https://cerinamroth.com/research/receipts-driven-engineering/"
    },
    {
      "text": "Little red dots — an ongoing search Map published little red dots, explore an ongoing search of public JWST images, and follow individual sources through real telescope observations.",
      "url": "https://cerinamroth.com/jwst/lrds/"
    },
    {
      "text": "Vulnerability research Good-faith research & coordinated disclosure.",
      "url": "https://cerinamroth.com/policy/"
    },
    {
      "text": "Published disclosures ZisK zkVM: a soundness bug in signed remainder, fixed upstream. Vendor advisory GHSA-qjgg-fcj3-p9x3, rated Critical.",
      "url": "https://cerinamroth.com/disclosures/"
    },
    {
      "text": "jordan@cerinamroth.com",
      "url": "mailto:jordan@cerinamroth.com"
    },
    {
      "text": "security@cerinamroth.com",
      "url": "mailto:security@cerinamroth.com"
    },
    {
      "text": "security.txt",
      "url": "https://cerinamroth.com/.well-known/security.txt"
    },
    {
      "text": "jordananderson.work",
      "url": "https://jordananderson.work"
    },
    {
      "text": "Site data",
      "url": "https://cerinamroth.com/site.json"
    },
    {
      "text": "public GPG key",
      "url": "https://cerinamroth.com/.well-known/jordananderson-pubkey.asc"
    }
  ],
  "structured_data": [
    {
      "@context": "https://schema.org",
      "@graph": [
        {
          "@type": "Organization",
          "@id": "https://cerinamroth.com/#organization",
          "name": "Cerin Amroth Research",
          "legalName": "Cerin Amroth Security Research, LLC",
          "url": "https://cerinamroth.com/",
          "logo": "https://cerinamroth.com/brand/tree-icon-180-33ec12628491.png",
          "description": "Independent research in 4-bit Mixture-of-Experts models, JWST little red dots and software security. Explore open-source tools, telescope images and evidence."
        },
        {
          "@type": "WebSite",
          "@id": "https://cerinamroth.com/#website",
          "url": "https://cerinamroth.com/",
          "name": "Cerin Amroth",
          "alternateName": "Cerin Amroth Research",
          "publisher": {
            "@id": "https://cerinamroth.com/#organization"
          }
        }
      ]
    }
  ],
  "content_markdown": "[Home](https://cerinamroth.com/)[ML](https://cerinamroth.com/ml/)[JWST](https://cerinamroth.com/jwst/)[Security](https://cerinamroth.com/policy/)[Contact](https://cerinamroth.com/#contact) \n\n# Cerin Amroth — independent research: ML-systems efficiency, JWST observational work, and software security\n\n      \n\nLarge models on smaller machines. Public telescope images, examined closely. Software claims, put to the test.\n\n  \n\n \n\n## ML systems — large models, smaller machines\n\n \n\n [ Loggetta Check your hardware, plan a training run, and keep the results. One install includes the runtime and kernels. v0.7.0 · Plan and run training · `pip install loggetta` → ](https://cerinamroth.com/ml/) [ experts4bit-qlora Fine-tune and serve large MoE models on one GPU. Move weights to RAM or SSD when GPU memory runs out. v0.58.1 · Get started → ](https://cerinamroth.com/ml/experts4bit-qlora/) [ grouped-nf4-gemm Compute directly on packed 4-bit weights. GPU kernels, memory tools, and tests you can reproduce. v0.51.0 · Get started → ](https://cerinamroth.com/ml/grouped-nf4-gemm/) [ OpenVINO Upstream fixes keeping Intel-GPU LLM inference off the CPU-fallback path — three merged. ](https://cerinamroth.com/ml/openvino/) \n\n \n\n[Latest ML research · 8 October 2026](https://cerinamroth.com/research/#update-2026-10-08) · [GitHub: Loggetta](https://github.com/pjordanandrsn/loggetta) · [PyPI](https://pypi.org/project/loggetta/) · [models](https://pypi.org/project/loggetta/#models) · [measured speed and architecture](https://cerinamroth.com/ml/) · [lower-layer solution guides](https://cerinamroth.com/ml/solutions/) \n\n \n\n \n\n \n\n## Start with the problem\n\n \n\n | Problem | Route \n | I want to plan and run model training | [Loggetta](https://cerinamroth.com/ml/) \n | My MoE still OOMs after `load_in_4bit=True` | [experts4bit-qlora internals](https://cerinamroth.com/ml/solutions/bitsandbytes-moe-load-in-4bit-still-ooms/) \n | I need to QLoRA-train fused MoE experts | [experts4bit-qlora](https://cerinamroth.com/ml/solutions/qlora-fused-moe-experts/) \n | My model is larger than GPU or host memory | [e4b residency paths](https://cerinamroth.com/ml/solutions/run-moe-larger-than-vram/) \n | Dequantizing every expert before GEMM is too slow | [grouped-nf4-gemm](https://cerinamroth.com/ml/solutions/nf4-grouped-gemm-without-bf16-materialization/) \n | I need native MXFP4 or INT4 expert compute | [grouped-nf4-gemm](https://cerinamroth.com/ml/solutions/native-mxfp4-moe-inference/) \n | I want to inspect the evidence methodology | [Receipts-driven engineering](https://cerinamroth.com/research/receipts-driven-engineering/) \n\n \n\n \n\n \n\n## Astronomy — explore real JWST observations\n\n \n\n [ Little red dots — an ongoing search Map published little red dots, explore an ongoing search of public JWST images, and follow individual sources through real telescope observations. ](https://cerinamroth.com/jwst/lrds/) \n\n \n\n \n\n \n\n## Security — find the flaw, help get it fixed\n\n \n\n [ Vulnerability research Good-faith research & coordinated disclosure. ](https://cerinamroth.com/policy/) [ Published disclosures ZisK zkVM: a soundness bug in signed remainder, fixed upstream. Vendor advisory GHSA-qjgg-fcj3-p9x3, rated Critical. ](https://cerinamroth.com/disclosures/) \n\n \n\n  \n\nContact: [jordan@cerinamroth.com](mailto:jordan@cerinamroth.com) · security reports: [security@cerinamroth.com](mailto:security@cerinamroth.com) · [security.txt](https://cerinamroth.com/.well-known/security.txt) · consulting: [jordananderson.work](https://jordananderson.work) · [Site data](https://cerinamroth.com/site.json)\n\n \n\nJordan Anderson’s [public GPG key](https://cerinamroth.com/.well-known/jordananderson-pubkey.asc) · Ed25519, expires 11 June 2028\nFingerprint: `DAEF C03B 9188 793C 80E5 20C8 68F0 6663 C7FB 250F`\n\n  © 2026 Cerin Amroth Security Research, LLC · CERIN AMROTH SECURITY RESEARCH™\n",
  "content_sha256": "307d2c94679d82c7cd0a443f0202d7924dc961c21727f02bcb34cc6bcdfae602"
}