{
  "version": "https://jsonfeed.org/version/1.1",
  "title": "Flexcompute Engineering",
  "home_page_url": "https://engineering.flexcompute.com",
  "feed_url": "https://engineering.flexcompute.com/feed.json",
  "description": "Essays, tutorials, and case studies on AI engineering, computational physics, photonics, and simulation from Flexcompute.",
  "items": [
    {
      "id": "https://engineering.flexcompute.com/articles/why-simulate-the-world/",
      "url": "https://engineering.flexcompute.com/articles/why-simulate-the-world/",
      "title": "Why Simulate the World",
      "summary": "When thinking and building become cheap, proof from the real world becomes the slowest step in progress. A second world, built from the laws of physics, makes proving fast: machines test designs there and build only the winners.",
      "image": "https://engineering.flexcompute.com/images/og/why-simulate-the-world.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/why-simulate-the-world.png",
      "date_published": "2026-08-28T00:00:00.000Z",
      "authors": [
        {
          "name": "Zongfu Yu",
          "url": "https://engineering.flexcompute.com/authors/zongfu-yu/"
        }
      ],
      "tags": [],
      "content_text": "When thinking and building become cheap, proof from the real world becomes the slowest step in progress. A second world, built from the laws of physics, makes proving fast: machines test designs there and build only the winners.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/why-simulate-the-world.md",
          "mime_type": "text/markdown",
          "title": "Why Simulate the World markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/simulating-the-world/",
      "url": "https://engineering.flexcompute.com/articles/simulating-the-world/",
      "title": "Simulating the World",
      "summary": "Language predicts a storm without resolving a single molecule. First-principles simulation resolves every molecule but cannot reach a single living cell. The way to simulate the world is a new tokenization of physics.",
      "image": "https://engineering.flexcompute.com/images/og/simulating-the-world.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/simulating-the-world.png",
      "date_published": "2026-08-24T00:00:00.000Z",
      "authors": [
        {
          "name": "Zongfu Yu",
          "url": "https://engineering.flexcompute.com/authors/zongfu-yu/"
        }
      ],
      "tags": [],
      "content_text": "Language predicts a storm without resolving a single molecule. First-principles simulation resolves every molecule but cannot reach a single living cell. The way to simulate the world is a new tokenization of physics.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/simulating-the-world.md",
          "mime_type": "text/markdown",
          "title": "Simulating the World markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/a-physics-company/",
      "url": "https://engineering.flexcompute.com/articles/a-physics-company/",
      "title": "A Physics Company",
      "summary": "The laws of physics have been known for a century. What those laws do, at scale, has never been computed — and that gap is where an entire industry is still waiting to be built.",
      "image": "https://engineering.flexcompute.com/images/og/a-physics-company.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/a-physics-company.png",
      "date_published": "2026-08-18T00:00:00.000Z",
      "authors": [
        {
          "name": "Zongfu Yu",
          "url": "https://engineering.flexcompute.com/authors/zongfu-yu/"
        }
      ],
      "tags": [],
      "content_text": "The laws of physics have been known for a century. What those laws do, at scale, has never been computed — and that gap is where an entire industry is still waiting to be built.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/a-physics-company.md",
          "mime_type": "text/markdown",
          "title": "A Physics Company markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/liveviewer-loop-engineering/",
      "url": "https://engineering.flexcompute.com/articles/liveviewer-loop-engineering/",
      "title": "We built a layout viewer for AI agents. Then an agent scaled it to 38 million shapes.",
      "summary": "How a goal-driven loop turned PhotonForge's browser LiveViewer from a half-million-shape viewer into a 38-million-shape, 60 FPS tool for humans and agents.",
      "image": "https://engineering.flexcompute.com/images/og/liveviewer-loop.jpg",
      "banner_image": "https://engineering.flexcompute.com/images/og/liveviewer-loop.jpg",
      "date_published": "2026-06-25T00:00:00.000Z",
      "authors": [
        {
          "name": "Prash Kharel",
          "url": "https://engineering.flexcompute.com/authors/prash-kharel/"
        },
        {
          "name": "Darcy Parker",
          "url": "https://engineering.flexcompute.com/authors/darcy-parker/"
        }
      ],
      "tags": [
        "AI Engineering",
        "AI Agents",
        "Photonics",
        "Verification"
      ],
      "content_text": "How a goal-driven loop turned PhotonForge's browser LiveViewer from a half-million-shape viewer into a 38-million-shape, 60 FPS tool for humans and agents.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/liveviewer-loop-engineering.md",
          "mime_type": "text/markdown",
          "title": "We built a layout viewer for AI agents. Then an agent scaled it to 38 million shapes. markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/electrical-routing-agents/",
      "url": "https://engineering.flexcompute.com/articles/electrical-routing-agents/",
      "title": "Learning Auto-Routing by Building: From Brute Force to an Auto-Design Agent",
      "summary": "How I learned auto-routing for photonic chips — a failed brute-force attempt, AI as a learning partner, interactive HTML arenas, a PhotonForge router, and an agent that iterates 27 designs in under three minutes.",
      "image": "https://engineering.flexcompute.com/images/og/electrical-routing.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/electrical-routing.png",
      "date_published": "2026-04-20T00:00:00.000Z",
      "authors": [
        {
          "name": "Prash Kharel",
          "url": "https://engineering.flexcompute.com/authors/prash-kharel/"
        }
      ],
      "tags": [
        "AI Engineering",
        "AI Agents",
        "Photonics"
      ],
      "content_text": "How I learned auto-routing for photonic chips — a failed brute-force attempt, AI as a learning partner, interactive HTML arenas, a PhotonForge router, and an agent that iterates 27 designs in under three minutes.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/electrical-routing-agents.md",
          "mime_type": "text/markdown",
          "title": "Learning Auto-Routing by Building: From Brute Force to an Auto-Design Agent markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/mode-solver-memory-calibration/",
      "url": "https://engineering.flexcompute.com/articles/mode-solver-memory-calibration/",
      "title": "Predicting Peak Memory for an Electromagnetic Mode Solver",
      "summary": "How we replaced a heuristic memory estimate with a calibrated model for Tidy3D mode solver workloads, eliminating under-predictions across the calibration set.",
      "image": "https://engineering.flexcompute.com/images/og/mode-solver-memory-calibration.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/mode-solver-memory-calibration.png",
      "date_published": "2026-04-15T00:00:00.000Z",
      "authors": [
        {
          "name": "Momchil Minkov",
          "url": "https://engineering.flexcompute.com/authors/momchil-minkov/"
        }
      ],
      "tags": [
        "Photonics",
        "Tidy3D",
        "Verification"
      ],
      "content_text": "How we replaced a heuristic memory estimate with a calibrated model for Tidy3D mode solver workloads, eliminating under-predictions across the calibration set.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/mode-solver-memory-calibration.md",
          "mime_type": "text/markdown",
          "title": "Predicting Peak Memory for an Electromagnetic Mode Solver markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/autoresearch-photonic-design/",
      "url": "https://engineering.flexcompute.com/articles/autoresearch-photonic-design/",
      "title": "Can AI Agents Autonomously Design Components on Photonic Chips?",
      "summary": "We gave AI agents a photonic simulator, a DRC engine, and four design challenges. They autonomously designed waveguide bends, crossings, splitters, and demultiplexers — some reaching near-perfect performance.",
      "image": "https://engineering.flexcompute.com/images/og/autoresearch-photonic-design.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/autoresearch-photonic-design.png",
      "date_published": "2026-04-13T00:00:00.000Z",
      "authors": [
        {
          "name": "Tyler Hughes",
          "url": "https://engineering.flexcompute.com/authors/tyler-hughes/"
        }
      ],
      "tags": [
        "Photonics",
        "Inverse Design",
        "Optimization",
        "AI Agents"
      ],
      "content_text": "We gave AI agents a photonic simulator, a DRC engine, and four design challenges. They autonomously designed waveguide bends, crossings, splitters, and demultiplexers — some reaching near-perfect performance.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/autoresearch-photonic-design.md",
          "mime_type": "text/markdown",
          "title": "Can AI Agents Autonomously Design Components on Photonic Chips? markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/photonic-inverse-design-45-lines/",
      "url": "https://engineering.flexcompute.com/articles/photonic-inverse-design-45-lines/",
      "title": "Designing a Photonic Chip Component with ~45 Lines of Python",
      "summary": "A compact introduction to photonic inverse design with Tidy3D, using a pre-built simulation and a ~45-line optimization loop.",
      "image": "https://engineering.flexcompute.com/images/og/photonic-inverse-design-45-lines.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/photonic-inverse-design-45-lines.png",
      "date_published": "2026-03-05T00:00:00.000Z",
      "authors": [
        {
          "name": "Tyler Hughes",
          "url": "https://engineering.flexcompute.com/authors/tyler-hughes/"
        }
      ],
      "tags": [
        "Photonics",
        "Inverse Design",
        "Optimization",
        "Tidy3D"
      ],
      "content_text": "A compact introduction to photonic inverse design with Tidy3D, using a pre-built simulation and a ~45-line optimization loop.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/photonic-inverse-design-45-lines.md",
          "mime_type": "text/markdown",
          "title": "Designing a Photonic Chip Component with ~45 Lines of Python markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/what-should-we-work-on-next/",
      "url": "https://engineering.flexcompute.com/articles/what-should-we-work-on-next/",
      "title": "\"What Should We Work On Next?\"",
      "summary": "The story of building an 80,000-line autodiff library almost entirely through AI agents — and the verification infrastructure that made it possible.",
      "image": "https://engineering.flexcompute.com/images/og/what-should-we-work-on-next.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/what-should-we-work-on-next.png",
      "date_published": "2026-02-26T00:00:00.000Z",
      "authors": [
        {
          "name": "Yannick Augenstein",
          "url": "https://engineering.flexcompute.com/authors/yannick-augenstein/"
        },
        {
          "name": "Frederik Schubert",
          "url": "https://engineering.flexcompute.com/authors/frederik-schubert/"
        }
      ],
      "tags": [
        "AI Engineering",
        "Autodiff",
        "Verification"
      ],
      "content_text": "The story of building an 80,000-line autodiff library almost entirely through AI agents — and the verification infrastructure that made it possible.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/what-should-we-work-on-next.md",
          "mime_type": "text/markdown",
          "title": "\"What Should We Work On Next?\" markdown"
        }
      ]
    },
    {
      "id": "https://engineering.flexcompute.com/articles/agent-control-loop/",
      "url": "https://engineering.flexcompute.com/articles/agent-control-loop/",
      "title": "The Agent Control Loop — Engineering for Tolerance",
      "summary": "Why agent reliability isn't magic model behavior — it's an environment where correctness is continuously verified. A framework for deciding when and how to delegate to AI agents.",
      "image": "https://engineering.flexcompute.com/images/og/agent-control-loop.png",
      "banner_image": "https://engineering.flexcompute.com/images/og/agent-control-loop.png",
      "date_published": "2026-01-19T00:00:00.000Z",
      "authors": [
        {
          "name": "Frederik Schubert",
          "url": "https://engineering.flexcompute.com/authors/frederik-schubert/"
        },
        {
          "name": "Yannick Augenstein",
          "url": "https://engineering.flexcompute.com/authors/yannick-augenstein/"
        }
      ],
      "tags": [
        "AI Engineering",
        "AI Agents",
        "Verification"
      ],
      "content_text": "Why agent reliability isn't magic model behavior — it's an environment where correctness is continuously verified. A framework for deciding when and how to delegate to AI agents.",
      "attachments": [
        {
          "url": "https://engineering.flexcompute.com/articles/agent-control-loop.md",
          "mime_type": "text/markdown",
          "title": "The Agent Control Loop — Engineering for Tolerance markdown"
        }
      ]
    }
  ]
}