I like computers most when they have something to prove.
I build around zkVMs, rollups, privacy, and distributed systems. Most things here start with a fairly simple question: can this execution be proven, can this system be observed, or can something private be verified without exposing it?
Then I write enough code to find out.
Current rabbit holes include proving OP Stack execution, running Solana BPF inside a zkVM, making long-running prover jobs easier to inspect, implementing cryptographic primitives and precompiles, and producing verifiable receipts for private agent context.
Some projects become useful tools. Some remain rough experiments. Some are evidence that the original question was much harder than it looked.
I care about software that is verifiable, private when it needs to be, and usable without pretending the machinery underneath does not exist.
Mostly building, breaking, and documenting things. Occasionally in that order.


