We're an applied research lab working towards a future where engineers can use natural language to do CAD tasks end-to-end. With high-level commands like "Hey Camfer, design a gearbox w/ a 3:1 ratio for this motor!", our agents will autonomously design the system, source off-the-shelf parts, and create drawings to be sent off for manufacturing (like Jarvis for Ironman). Last fall, we won both Y Combinator's internal demo day and OpenAI's o1 hackathon! Here's a clip of Sam Altman talking about Camfer.

Our first step is creating a tool that allows engineers to use text, images, and mechanical drawings to generate feature-rich CAD models directly in SolidWorks and Onshape. We've got a few pilot users at companies like Physical Intelligence using our v0.

We're a small, young team of software devs and researchers who have a love for CAD and design. Our founders are FRC world champions, UC Berkeley alumni / researchers, ex-AWS, and ex-Meta.

We finished YC in October and raised a $4.8M seed round. Our investors include Y Combinator, Next47, Bloomberg Beta, Octave Fund, and the founders of Dropbox, Airtable and Clari.

We're in-person in SF. We have an espresso machine, meals provided every day, a fully stocked fridge, and lots of natural light!

Working at Camfer

At Camfer, interns are given the same autonomy as full-time engineers and researchers. Native Sparse Attention was an intern project! The only bottlenecks on your impact will be your imagination, taste, and agency.

We’re building a flat and talent dense organization. We enjoy deep, multi-disciplinary thinking, high-conviction bets, spirited debate, and having fun!

Role

We are looking for people who:

• Are generalists that have worked with a broad range of programming languages and systems
• Are passionate, creative, curious, and truth-seeking.
• Have the ability to adapt quickly and are always learning

Some example projects you might work on:

• Training reasoning models with RLVR to fully constrain CAD sketches
• Developing new AI-native UI components for the future of CAD
• Adapting our training pipelines to new modalities like point clouds
• Expanding platform integrations based on customer needs
• Optimizing model hosting/inference for faster and more robust generations