AI Engineering Intern

Work with me on the engine that reads CAD models and technical drawings and turns them into machining plans.

KinetiCAD reads a part's CAD model and its technical drawing and works out how a machine shop should make it. It picks the machine, the tools and the operations, and it writes the CNC program. Every program is checked against the CAD model in a cutting simulation before it can be released. We're based at the Berkeley Sky Deck accelerator.

You'll work on the engine with me. I lead the geometry and AI side. Every problem starts with a part and its drawing, and the goal is a plan a machinist would sign off on. Interns and junior engineers write production code here, and you'll see yours running on parts in your first weeks.

• Make the engine better at reading technical drawings, including dimensions, tolerances, surface finishes and notes.
• Work on finding machining features in 3D CAD models and turning them into operations.
• Build out the evaluation sets. That means collecting parts and expected results, and running accuracy checks so we know whether a change helped or hurt.
• Work with the local language models in the product, and know when a plain deterministic algorithm does the job better.
• Write tests, read profiles and speed up whatever is slow.

• Strong Python. You've built something with it, at school, at work or on your own.
• Machine learning fundamentals and hands-on time with PyTorch or a similar framework.
• You're studying computer science, mechanical or electrical engineering, physics or maths, or you just graduated.
• You're curious about how physical parts get made.

• 3D or computational geometry, or a geometry kernel like Open CASCADE.
• Computer vision or document understanding (OCR, layout, reading drawings).
• You've run language models locally.
• Any time spent in a machine shop or using CAD software.