Autonomous market infrastructure
Aurifern builds autonomous systems for on-chain markets, and measures every one of them long before it touches money.
Method
Every result keeps its inputs, outputs, timing and code version.
Limits on time, spend and failure are set before anything runs.
Replay, monitoring and rollback carry over to the next engine.
Work
A few engines and the ground they share. Small on purpose.
Flagship
Batch auctions reward the solver that finds the best settlement for everyone in the batch. This engine competes for that, carefully: every answer is simulated first, and every change is tested against history before it goes anywhere near live flow.
Shared plumbing that makes every engine checkable: what it saw, what it did, which version did it, and how to replay the moment exactly. Every engine runs on it.
A read-only lane for studying intent-based order flow: which orders appear, how they resolve, and what a filler could have done. It watches and records. Nothing more.
New engines enter the same way: research, simulation, a capital cap, a gate.
Architecture
Chain data, event feeds and endpoint health. Every later claim points back to a record kept here.
Rebuilds past opportunities from saved inputs and replays them, including what the engine would have done.
Where signing, nonce handling, simulate-before-send and loss limits live, walled off from everything else.
One adapter per protocol. Each keeps its own risk book, its own gate and its own evidence.
Status, incidents and metrics taken from what actually ran. No forecasts.
About
Aurifern is founded and run by Danil Zak. Automation does the repetitive work. Judgement stays with a person. Nothing touches capital before it has been measured.
Danil ZakFounder
Say hello
Building something where correctness matters more than volume? Write to us. Small experiments first.