Product leader focused on execution, reliability, and outcomes at scale
Self-improving AI agent trading battle
Problem
AI agents operate invisibly. There's no way to watch them work, compare their strategies, or understand how they make decisions in real time.
What I did
Built a live PvP arena where two autonomous agents trade real money, evolving their own strategy every 2 hours trying to beat each other.
Multichain tokenization infrastructure
Problem
Token issuance across chains was fragmented, making execution reliability hard to achieve.
What I did
Defined token lifecycle primitives and execution guarantees, prioritizing predictable behavior and recovery paths over flexibility.
Impact
Deployed in production across 9 networks, processing 15.9M transactions, including 2.73M cross-chain transactions.
Primary execution interface for Holograph Protocol
Problem
Users lacked clear visibility into transaction state, leading to hesitation, failed retries, and abandonment.
What I did
Redesigned execution flows around explicit transaction state, automated recovery for stalled transactions, and clear post-failure feedback.
Impact
Supported 11.46M mints across 3.19M unique wallets.
Trading and asset discovery interface on Base
Problem
Users struggled to discover assets and act confidently due to fragmented market data and unclear execution context.
What I did
Designed fast, legible market surfaces that prioritized execution clarity, asset discovery, and ecosystem-specific context.
Impact
Enabled multi-asset trading flows, improved execution confidence, and increased repeat usage across active traders.
Mobile memecoin trading app on Base
Problem
Onchain trading had too much friction before first trade. I reduced time-to-first-trade to test whether UX alone could unlock sustained demand on Base.
What I did
Designed onboarding and execution flows optimized for speed, including email-based signup and instant trades, to compress time-to-first-trade and reduce cognitive overhead.
Impact
Shortened time to first trade, increased first-session completion, and supported repeat trading behavior among new users in fast-moving markets.
Daily competitive prediction game for memecoin markets
Why: Test whether short-cycle competition and social sharing could drive engagement beyond buy-and-hold behavior.
What it showed: Competition drove bursts of activity, but retention depended on narrative momentum.
Creator coin with automated buyback and burn mechanics
Why: Test whether programmatic supply reduction can create durable incentives beyond hype cycles.
What it showed: Self-reinforcing loops work onchain, but only with sufficient liquidity and sustained activity.
One-shot AI tool for high-stakes product decisions
Why: Test whether structured decision framing can outperform chat-based ideation for irreversible product choices.
What it showed: Constraint-driven framing produces clearer, more defensible decision artifacts than open-ended brainstorming.
CLI tool for detecting behavioral changes in Base wallet addresses
Why: Test whether observable interaction diffs can serve as trust signals without relying on subjective safety scores.
What it showed: Change detection is more defensible than risk verdicts, but baseline quality and RPC limits determine whether users trust the signal.
I'm Jason (aka Lost Midas). I'm a product leader with 9 years of experience building crypto infrastructure and trading systems. Most of my work was at Holograph, where I led product for a multichain tokenization protocol that handled 15.9M transactions across 9 networks. I focus on execution problems: how systems behave under load, how users recover from failures, how incentives shape behavior at scale. The products I care about are the ones where correctness matters more than features, and where real capital and user trust are at stake.