Raychis

Field notes

Building Raychis in the open.

A 10-part series on building a real-world, on-device plant identifier from scratch, written while it was still uncertain whether it would work. Published a part a week, with the last landing the day the app went live.

The complete series: all 10 parts published

  1. Part 1 of 10

    Plant Identification Looks Solved. Until You Try to Build It

    An app told me a toxic plant was basil. I spent a year building a plant identifier from scratch to understand why, and to find where computer vision quietly breaks.

  2. Part 2 of 10

    The First Thing That Broke Was the Data

    Before the model, the data. Corrupt images, mislabelled species, and the unglamorous pipeline work that decides whether anything downstream can work at all.

  3. Part 3 of 10

    The Data Looked Fixed. The Model Knew Better

    The class distribution finally looked right, and the model still failed. What clean-looking data hides, and how training surfaces the problems you thought you had solved.

  4. Part 4 of 10

    The Commit Before the Silence

    The last commit of 2025, then fifty-two days of nothing. What stalls a solo build, and what it takes to come back to a project you have walked away from.

  5. Part 5 of 10

    The Two Weeks I Stopped Building

    Coming back to the project, getting the pipeline running end to end, and putting out the fire that had nearly killed it in November.

  6. Part 6 of 10

    The Data Valley

    The first real model: 42.7% top-1 on 2,592 species, carefully logged and decisively unshippable. Sitting in the trough where the numbers say you are nowhere near done.

  7. Part 7 of 10

    The Reframe

    Reading papers in the valley, and the change in thinking that moved the project forward when brute force had stopped working.

  8. Part 8 of 10

    The Breakage

    I went on holiday on April 3rd. What broke while I was away, and what that revealed about the system I thought I had built.

  9. Part 9 of 10

    The Rule I Set for Myself: The Stage Gate

    Accuracy ticking up, the clock ticking down, and the rule that kept me honest about what was actually ready to ship.

  10. Part 10 of 10

    The App That Will Tell You It Might Be Wrong

    Nine months after a plant app called annual mercury basil, an app that says when it is not sure. Where the series began, and where it lands.

The series is published first on Medium; these are the same pieces, kept here for the record.