Kseniia Shevchenko

Generation tools are low-stakes when they're wrong draft an email, get a bad draft, edit it. Judgment tools are different: grading, flagging, scoring. If the output looks confident and it's actually a guess, someone might act on it without knowing that.

Working on letsflw's exam-checker made this concrete objective questions are easy (exact match, no ambiguity), but for open-ended answers we made a call to surface the AI's mark as a suggestion to review, not a final score, because presenting it as settled felt dishonest.

Before opening a PR, I started pasting my code into Claude for a quick review first, just to catch anything obvious before it goes to the lead.

A few times it caught things clearly worth fixing: a missed null case, a variable name that didn't match the rest of the file's convention, a catch block that swallowed the error without logging it. Stuff that, if the lead caught it instead, would've meant another round of comment - fix - push.

But there were also times it gave generic "best practice" suggestions that didn't fit our team's actual conventions, so I had to filter through what to actually apply versus ignore.

Not always a time-saver either, for small PRs, the lead's review still came back basically unchanged after I'd already run it through Claude, so that step felt redundant. The benefit was clearer on PRs with more complex logic.

Hey everyone,

I've been building a labelling tool, a desktop app (PyQt6) for capturing and annotating gesture datasets think of it as a companion tool for gesture recognition pipelines like Gesto, making it easier to collect, label, and export hand landmark data for training models.

I'm at a fork in the road on direction, and I'd love your input:

1. Open source it release it freely (currently GPL-3.0), let the community use it, contribute, and shape where it goes.