Partner at NGC Ventures
Check size: No personal or firm ticket is disclosed on NGC's site. Zi's partner title, accelerator mentorship and infrastructure responsibilities show meaningful technical involvement but do not prove a standard range or unilateral decision authority.
Zi focuses on blockchain infrastructure from research through live network operation. He evaluates execution bottlenecks, layer-2 and appchain architecture, EVM compatibility and parallel processing, while StakeX exposes him to validator operations across networks such as NYM, Algorand, SKALE and Avalanche. The result is a practical lens on whether protocol performance claims survive testnets, incentives and production operations.
Bring the architecture and benchmark harness. Explain the current bottleneck, execution model, sequencing, compatibility, security assumptions and results under realistic load. Show testnet partners, validator requirements, developer migration effort and economics. For accelerator-stage teams, identify the next technical milestone and evidence NGC can help validate.
His published Smooth Labs analysis favors teams that identify architectural bottlenecks early, combine complementary founder skills, refine a technical approach over time and can demonstrate a material performance improvement without abandoning existing developer compatibility. Operational credibility on testnets and validators matters.
Inferred from his work: throughput claims without reproducible tests, scaling approaches that simply move congestion elsewhere, architectures that sacrifice compatibility without sufficient benefit, weak validator economics, and infrastructure teams without the patience or operational skill to move from design to testnet and mainnet.
Measure the current virtual-machine or sequencing bottleneck under realistic load, then compare throughput, latency and failure behavior before and after the proposed parallelization.
Judge performance gains alongside developer migration cost, EVM/tooling compatibility, composability and security rather than maximizing transactions per second in isolation.
Validate an infrastructure thesis through testnet deployment, node operations, incentive behavior and mainnet reliability so research claims are tested against production constraints.
Parallel execution middleware can relieve L2 and appchain congestion while preserving EVM compatibility, provided a technically complementary team proves performance on real testnets.