Kadair Eskandar

Hi, I'm Kadair! I'm building Catalyst, a residency that turns academic researchers into founders.

I've run operations for two ML researcher communities, AGI House SF and DEMi, and I'm alarmed by how frontier AI research is concentrating in a few closed labs, leaving very few researchers able to participate.

Many promising ideas from brilliant academic researchers fade in the lab before they see light in the market. The AI market will be worth tens of trillions, and more builders must participate. There is plenty of untapped founder talent in academic research, and I'm creating a mechanism for researchers to test their ideas, apply their research, and start frontier research companies.

In the past

• I studied biomedical engineering and biodesign at Washington University in Saint Louis
• In grad school, I designed a new method to access arteries during high risk vascular surgeries with Dr. Mohamed Zayed
• I later joined a medical robotic startup and was mentored by Frank Grillo, where I researched and helped develop tomography as a method to accurately locate lung tumors in quasi-realtime during biopsy. We went from a prototype at the time I joined to FDA clearance in less than 2 years, and now the robot has treated over 10,000 patients.
• In 2023, I joined Jeremy Nixon to start AGI House SF and ran it as head of operations - it quickly became a central node for ML researchers and hackers to build together and start companies. Many AI companies that did well had a touchpoint with the house like Perplexity, Replicate, AGI Inc., Periodic Labs
• I later joined Kyle Morris to help run DEMi as head of community operations. I ran 2 conferences, and several hackathons, workshops and dinners, and helped grow the network to hundreds of researchers, engineers, and policymakers.

As for research, I'm personally drawn to society simulation and AI-AI interactions, particularly Serina Chang's (took her ML & Human Behavior course at Berkeley) and Joon Sung's work, to understand how these systems will reshape human coordination at scale.

As I reflect on my work, one consistent thread: I find it deeply rewarding to 1. pick and support highly agentic and ambitious builders who have a strong theory for how to shape the future and 2. design an environment for them to do their best work.

If you're an ML academic researcher interested in applying your research, reach out at kadair@demi.so.