Home›Companies›Dydema
Dydema

A consumer brain-computer interface you can map to anything.

We’re building a brain-computer interface anyone can wear, so you can use a computer at the speed of thought. To get there, we’re recording brain data at model company scale.
Active Founders
Julian Weaver
Julian Weaver
Founder/CEO
Founder & CEO at Dydema. We’re building personal, programmable brain-computer interfaces. Prev. founding engineer at Syntra (YC S24) and co-founder of Longhorn Neurotech (200+ members). Led EEG/fMRI AI research; dual BS in CS and neuroscience @ age 20, UT Austin.
Founder and CTO at Dydema. Previously: Founding Software Engineer at Syntra, ML researcher at UH SAIL, and Studied CS at University of Houston.
Company Launches
Dydema: Personal, programmable brain-computer interfaces
See original launch post

TL;DR: Dydema is building personal, programmable brain-computer interfaces- a headset plus the AI models behind it. You pick a thought and bind it to whatever you want, like a key in a game or a command to an agent. No surgery required. Our models already run on research headsets, and we’re designing our own consumer hardware.

https://youtu.be/zJCDVuodHzk

Hey y’all! We’re Julian Weaver and Christopher Oosthuizen, the founders of Dydema.

Brain-computer interfaces have mostly lived in labs. We want one you can order online, take out of the box, set up in a few minutes and use with any app on your computer, like a keyboard or a mouse. You decide which thoughts it listens for (imagined movements, for instance) and what each one does. The agent in the video is one example; it works the same way with games, tools, robots, and whatever else you hook it up to.

Where we are

  • After a few minutes of setup, our general decoding models work on someone they’ve never seen before, and on our benchmarks they beat models trained only on that person’s own data.
  • We’re pushing on how many distinct thoughts the models can be programmed to tell apart, and how fast they decode each one. Every new differentiable thought is another key on the keyboard.
  • We’re training the models to hold up on cheaper sensors and designing our headset around them.

We treat BCI as a machine learning problem, which mostly means needing lots of brain data. So we’re building the infrastructure to record more of our own. We’ve accumulated, cleaned and standardized ~14.6 years (128k hours) of EEG data, and we’re starting to record in real world contexts. People have debated for years about whether EEG models scale, and so far ours keep getting better with more labeled data.

Computers keep getting smarter, but the way we tell them what we want hasn’t changed in decades. Long term, we want a brain-computer interface that anyone can buy and use to communicate with computers and each other, faster than talking. We’ll keep posting demos as we go!

uploaded image

About us

  • Julian (CEO) graduated from UT Austin at 20 with degrees in CS and neuroscience. While there, he co-founded Longhorn Neurotech, which he grew into one of the largest student neurotech research orgs in the world (200+ members), and he led EEG/fMRI ML interpretability research.
  • Chris (CTO) was an ML researcher at UH SAIL, where he built state-of-the-art GraphRAG systems and worked on reducing LLM hallucinations. He’s been programming his whole life, and wrote his own programming language from scratch in high school.
  • We’ve been best friends since freshman year of high school, and we were both founding engineers at Syntra (YC S24).

Asks

  • Building something that could take a thought as an input (games, agents, robotics)? We’d love to talk.
  • If you have brain datasets, we want them!
  • We’ll be hiring more hardware engineers soon. If you know anyone who’s built wearables or low-noise biosignal hardware, please send them our way.
  • Want to try our rig? Let us know!

Cheers y’all,
Julian & Chris

Reach us at founders@dydema.com.

uploaded image

Dydema
Founded:2026
Batch:Fall 2026
Team Size:2
Status:
Active
Location:San Francisco
Primary Partner:Aaron Epstein