The latest articles on DEV Community by liesliy (@liesliy). https://dev.to/liesliy https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4057379%2F01a5a8ba-545c-4533-bd77-5f940e75f65c.png https://dev.to/liesliy en liesliy Sat, 01 Aug 2026 03:04:41 +0000 https://dev.to/liesliy/hello-devto-im-building-a-data-format-nobody-asked-for-but-everyone-needs-4a5i https://dev.to/liesliy/hello-devto-im-building-a-data-format-nobody-asked-for-but-everyone-needs-4a5i <p>Hey 👋 I'm Liesliy, a developer working on tactile data infrastructure for robots.<br> <strong>The Problem That Brought Me Here</strong><br> If you work in robotics, you've probably noticed something: nobody can agree on how to label tactile data.<br> Every lab has its own format. Every dataset ships in a different schema. When you try to combine data from two sources, you spend more time writing converters than doing actual research.<br> I've been there. So I built TLabel — an open-source unified format standard for tactile annotation data.</p> <p>Think of it like this:<br> ROS standardized how robots communicate<br> ROS Bag standardized how robots record<br> Nobody has standardized how we label tactile data</p> <p>That's the gap TLabel tries to fill.</p> <p><strong>What It Actually Does</strong><br> Unified schema with 14 semantic dimensions and 4 compliance levels (L1–L4)<br> Adapters to convert between formats (GelSight, BioTac, digit, TacTip, and more)<br> PyPI-installable, plug into your existing pipeline<br> It's early — v0.17 just shipped — but the architecture is solid and I'm actively looking for real-world feedback.</p> <p><strong>Why Dev.to</strong><br> I've been reading Dev.to for a while. What I appreciate about this place:<br> Engineers write for engineers — no fluff, no engagement farming<br> "Hello World" culture actually works — people genuinely welcome newcomers<br> The comment sections are better than most conferences<br> I'm here to:<br> Share what I learn building a data standard nobody asked for (yet)<br> Meet people working on sensor data, robotics, and embodied AI<br> Get honest feedback — the kind that hurts but makes the project better</p> <p><strong>What I'd Love to Hear From You</strong><br> If you work with tactile sensing, robotics data pipelines, or open-source tooling — I want to talk.<br> Especially curious:<br> How do you handle format conversion in your current workflow?<br> What's the most painful part of working with tactile datasets?</p> <p>Drop a comment or DM. I promise I'm more interested in listening than pitching. 🦞</p>