posts.json

[
  {
    "path": "posts/from-okd-to-k3s.html",
    "title": "Less Platform, More Research",
    "date": "2026-06-30",
    "tags": [
      "Engineering",
      "Kubernetes",
      "Open Source",
      "Infrastructure"
    ],
    "author": "Daniel Baldauf",
    "image": "static/posts/from-okd-to-k3s.png",
    "desc": "GPU drivers that required custom OS images, research software that fought Security Context Constraints, and 400 platform pods before a single workload scheduled. Why our research cluster moved from OKD to k3s."
  },
  {
    "path": "posts/ddevd-extreme-value-distribution.html",
    "title": "An Extreme Value Distribution That Knows Its Own Uncertainty",
    "titleDe": "Eine Extremwertverteilung, die ihre eigene Unsicherheit kennt",
    "date": "2026-06-29",
    "tags": [
      "Research",
      "Mathematics",
      "Statistics"
    ],
    "author": "Tobias Hell",
    "image": "static/posts/ddevd-extreme-value-distribution.png",
    "desc": "Two new papers introduce DDEVD, a non-parametric extreme value distribution that also provides a mathematically derived criterion for when its estimates can be trusted.",
    "descDe": "Zwei neue Arbeiten stellen DDEVD vor, eine nicht-parametrische Extremwertverteilung, die zugleich ein mathematisch hergeleitetes Kriterium liefert, wann ihrer Schätzung vertraut werden kann."
  },
  {
    "path": "posts/introducing-the-dlh-blog.html",
    "title": "Introducing the Data Lab Hell Blog",
    "titleDe": "Einführung in den Data Lab Hell Blog",
    "date": "2026-04-17",
    "tags": [
      "Company"
    ],
    "author": "Daniel Baldauf",
    "image": "static/brand/LOGO_DLH.svg",
    "desc": "We are a research and technology organisation and as an R&D Incubator & Innovation Lab unique in its form. Here is why we are starting to write publicly about our work.",
    "descDe": "Wir sind eine außeruniversitäre Forschungseinrichtung und R&D Incubator & Innovation Lab in Österreich. Hier erklären wir, warum wir beginnen, öffentlich über unsere Arbeit zu schreiben."
  },
  {
    "path": "posts/hacking-the-smart-home.html",
    "title": "Hacking the Smart Home",
    "titleDe": "Das Smart Home hacken",
    "date": "2026-04-20",
    "tags": [
      "Smart Home",
      "KNX",
      "Home Assistant",
      "Energy",
      "DIY"
    ],
    "author": "Johannes Sappl",
    "image": "static/posts/smart-home.png",
    "desc": "How I integrated KNX, Home Assistant, and dynamic EPEX Spot pricing to build a fully local, energy-aware smart home, from solar inverter to EV charging.",
    "descDe": "Wie ich KNX, Home Assistant und dynamische EPEX-Spot-Preise zu einem vollständig lokalen, energiebewussten Smart Home kombiniert habe, vom Wechselrichter bis zur E-Auto-Ladung."
  },
  {
    "path": "posts/ai-co-creation-mathematics.html",
    "title": "How we proved a structural rigidity theorem with an AI assistant, in a day",
    "date": "2026-05-27",
    "tags": [
      "Research",
      "AI",
      "Mathematics"
    ],
    "author": "Tobias Hell",
    "image": "static/posts/ai_cocreation.png",
    "desc": "A case study in human-AI co-creation in mathematics: how we proved a structural rigidity theorem about topological approaches to the Riemann Hypothesis, and what it means for interdisciplinary research."
  },
  {
    "path": "posts/our-open-source-projects.html",
    "title": "Our Open Source Projects",
    "date": "2026-06-24",
    "tags": [
      "Open Source",
      "Engineering",
      "Grafana",
      "Julia",
      "Kubernetes",
      "NVIDIA"
    ],
    "author": "Daniel Baldauf",
    "image": "static/posts/open-source-projects.png",
    "desc": "A look at some of the open-source tools we develop at Data Lab Hell (a backend-less Grafana PDF reporter, a one-file cgroups monitor, a binary ULID for Julia, and NVIDIA GPU images for OKD), plus where to find the rest."
  },
  {
    "path": "posts/faraday/index.html",
    "title": "Faraday: Scientific Context for AI",
    "date": "2026-07-03",
    "tags": [
      "MCP",
      "LLM",
      "Agentic AI",
      "Physics",
      "Chemistry",
      "Spectroscopy",
      "Research"
    ],
    "author": "Daniel Hauser",
    "image": "static/posts/faraday_logo.png",
    "desc": "Three end-to-end autonomous investigations (a fusion energy ledger, a methane flame's spectrum, and a molecule-by-molecule CO₂ vs CH₄ comparison), each built entirely on NIST and PubChem reference data retrieved through the open-source Faraday MCP server, with every figure traceable to a tool call."
  }
]