Two frontiers, drawn at the same time.
On the first, capital learns to keep and move itself on-chain. On the second, software stops answering and starts to act. The work here lives where each one is still unfinished, taking its newest primitive and making it something a careful person can put their weight on.
In crypto, that primitive is the vault: a programmable place for capital to live, be governed, and be put to work. In AI, it is the agent: a system raised on a house's own data and habits, that crosses from reply to action and can settle its own way when it must. Crypto is one room it runs in, not the whole house.
One discipline runs under both. Nothing ships on a claim a query cannot confirm. No agent is left alone with money before it has earned that trust. What follows is the work, and the thinking beneath it.
| Role | Crypto PM |
| Domain | Staking / DeFi |
| Also | Applied AI |
| Crypto | Vaults, validators |
| AI | Agents, RAG |
| Method | Query before claim |
| Status | Building |
What this does
Crypto vaults
The new primitive for capital on-chain: non-custodial, rules baked in, risk bounded. Staking loops, leverage, liquidation, carry. The SQL is written by hand, so every number has a query behind it.
AI agents
Autonomous agents trained on a business's own data and workflows: they act, not just answer, and can pay their own way when needed. The product work is control: guardrails, approvals, and evals on the whole trace. Crypto is one use case, not the limit.
Product, zero to one
Find the job nobody has done yet, then build the shortest path to it. Discovery, positioning, roadmap, and the unpopular call on what to cut.
Proof and compliance
Decisions settled by on-chain data, not decks. Regulation treated as a design input, not a tax paid at the end. That is often what lets a crypto product reach a serious institution.
Nothing ships on a claim a query cannot confirm. No agent is left alone with money before it has earned that trust.
Selected work
Leveraged Staking Carry Index
A live dashboard that answers one brutal question: at what leverage does looping staking still pay, and where does it blow up. Read off on-chain data, not vibes.
Leveraged Staking Vault
Turns the thin staking spread into something a fund can actually hold. The client gets the yield. They never touch a DeFi protocol.
Autonomous Agents and Machine Payments
Agents that act, fetch, and pay their own bills over x402. The product work is not the intelligence. It is the financial brakes that keep an autonomous spender from being insane.
Applied AI: RAG and Fine-tuning
Fine-tuning open models, mostly Gemma 4, on a single A100, and RAG over real databases. The unglamorous layer that decides whether an AI product is trustworthy or just a good demo.
Thoughts
Toolbelt
Take a break
Play a round ↗A round of mini-golf, hidden in here. A product is not finished until it is a little bit fun.