Research
Model a market mechanism, argue with it until it breaks, publish what you found — including when what you found was that it does not work.
What you’ll do
Take a hypothesis about market structure from a model to a simulated result to a reference implementation. Replay real chain state against candidate mechanisms and establish whether they survive contact with the order flow that actually exists, including its adversarial component.
Publish. Roughly half of what the lab publishes describes mechanisms we tested and rejected, and we consider that the more useful half. You will be expected to write, and to write clearly enough that a practitioner can disagree with you precisely.
What we look for
Depth in market microstructure, mechanism design, or applied probability — from academia, from a trading firm, or from having built something onchain that had to work. We are indifferent to which.
The ability to be specific about uncertainty. The work involves being wrong in public on a regular basis, and the people who do it well are the ones who state their confidence bounds before anyone asks.
Enough engineering fluency to build your own simulations. You do not need to ship production contracts; you do need to not be blocked waiting for someone who can.
How we work
Remote-first and asynchronous, with written proposals as the default unit of communication. Research questions are chosen through argument rather than assignment, and a well-argued case to kill a direction carries the same weight as a case to pursue one.
Applying
Write to Labs@kolkheti.org with something you have built, written or broken. We would rather read one thing you care about than a full CV.