Skip to content
View rd3n-ai's full-sized avatar

Block or report rd3n-ai

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
rd3n-ai/README.md

Nizam Yüksel — RD3N

Mechanical design engineer (SolidWorks, 11 years). I automate my own engineering work with Claude Code, and publish what actually ships.

I design custom industrial machines — roll-forming lines, presses, conveyors, special-purpose equipment — for manufacturers across 14+ sectors. Quoting, concept work, and CAD production used to eat my days. So I started rebuilding those steps around AI agents.

Everything here came out of paid client work, not tutorials. If a repo is public, it means the approach survived contact with a real machine shop.


What I build

The system I run my own shop on

hermes-server-kit — turn a spare mini PC into a private AI assistant for a machine shop, in one command. Firewall and SSH hardening, an agent you talk to over Telegram, and a git-backed company memory it maintains daily. Plain markdown, no vector database, no lock-in — your drawings and pricing stay on your own hardware.

I built it because my company's knowledge lived in my head. It now runs in a customer's factory too, where the owner uses it for payment tracking, project status, invoices, and drafting replies to email.

CAD, driven by code

onshape-api-belt-conveyor — an 82-part belt conveyor modeled entirely through the Onshape REST API, orchestrated by Claude Code. Live public Onshape document plus the real process behind it: what the API does well, where it fights you.

claude-code-mechanical-concept — a Claude Code skill that turns an underspecified equipment brief ("we need something that feeds sheet at 12 m/min") into a labeled spec table and a runnable CAD stub. Built because customers describe machines in fragments, and the missing numbers are the whole job.

From brief to something a customer can look at

machine-vision-tire-id-conveyor — warehouse-entry conveyor that reads tire brand, size, and DOT code off a moving wheel. Two numbers in a message, a 3D concept the same day.

draft-to-render-rollform-line — rough CAD used as the geometric anchor for an AI-generated factory render. The point: the render inherits real proportions instead of being a plausible-looking stock image.

3d-proposal-template — self-contained HTML proposal with an embedded, rotatable 3D model viewer. One file, no build step, no dependencies. A customer opens it on a phone in a workshop and rotates the machine you are quoting.


Why this combination

Most engineering-AI tooling is written by software people who then go learn manufacturing. I came the other way: I was specifying gear ratios and picking St52 plate long before I wrote a line of automation. That changes what you build — the failure modes I design around are wrong material callouts, lost revisions, and quotes that take three days, not abstract benchmarks.

I also run this on myself. My own shop knowledge — customers, machines, suppliers, decisions — lives in a git-based memory system that an agent reads and writes daily. The same structure is now installed in a customer's factory.


Open to

  • CAD and engineering automation, remote (Onshape/SolidWorks API work, generative geometry, drawing and BOM pipelines)
  • Claude Code skills and agent tooling for manufacturing workflows
  • Corporate memory / AI assistant setups for machine shops — currently on-site, Türkiye

Reach me through the email on my GitHub profile, or open an issue on any repo.

Popular repositories Loading

  1. hermes-server-kit hermes-server-kit Public

    Turn a spare mini PC into a private AI assistant for your machine shop — one command. Firewall, Telegram agent, and a git-backed company memory. Built by a mechanical engineer, running in a real fa…

    Shell 1

  2. onshape-api-belt-conveyor onshape-api-belt-conveyor Public

    82-part belt conveyor designed entirely through the Onshape API, driven by Claude Code — live public Onshape doc + the real process behind it.

    1

  3. machine-vision-tire-id-conveyor machine-vision-tire-id-conveyor Public

    Warehouse-entry conveyor concept that reads tire brand/size/DOT code off a moving wheel via machine vision — from a two-number brief to a 3D concept same day.

    1

  4. draft-to-render-rollform-line draft-to-render-rollform-line Public

    Rough CAD concept used as the visual reference for an AI-generated, quote-ready factory render — not a generic stock photo.

    1

  5. claude-code-mechanical-concept claude-code-mechanical-concept Public

    A Claude Code skill: turn an underspecified mechanical equipment brief into a labeled spec table + runnable CAD stub, fast.

    1

  6. 3d-proposal-template 3d-proposal-template Public

    Self-contained HTML proposal template with an embedded, rotatable 3D model viewer — for quoting equipment fast.

    HTML 1