Hey 👋 I'm Liesliy, a developer working on tactile data infrastructure for robots.
The Problem That Brought Me Here
If you work in robotics, you've probably noticed something: nobody can agree on how to label tactile data.
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.
I've been there. So I built TLabel — an open-source unified format standard for tactile annotation data.
Think of it like this:
ROS standardized how robots communicate
ROS Bag standardized how robots record
Nobody has standardized how we label tactile data
That's the gap TLabel tries to fill.
What It Actually Does
Unified schema with 14 semantic dimensions and 4 compliance levels (L1–L4)
Adapters to convert between formats (GelSight, BioTac, digit, TacTip, and more)
PyPI-installable, plug into your existing pipeline
It's early — v0.17 just shipped — but the architecture is solid and I'm actively looking for real-world feedback.
Why Dev.to
I've been reading Dev.to for a while. What I appreciate about this place:
Engineers write for engineers — no fluff, no engagement farming
"Hello World" culture actually works — people genuinely welcome newcomers
The comment sections are better than most conferences
I'm here to:
Share what I learn building a data standard nobody asked for (yet)
Meet people working on sensor data, robotics, and embodied AI
Get honest feedback — the kind that hurts but makes the project better
What I'd Love to Hear From You
If you work with tactile sensing, robotics data pipelines, or open-source tooling — I want to talk.
Especially curious:
How do you handle format conversion in your current workflow?
What's the most painful part of working with tactile datasets?
Drop a comment or DM. I promise I'm more interested in listening than pitching. 🦞
Top comments (0)