Overview

Uniopc is an AI business workspace built for real-economy industries, packaging the autonomous task planning and execution of AI agents into a product experience that one-person companies, small businesses, and lightweight operations teams can use directly — think of it as a WorkBuddy for real-world industries. The system is built around organizations, projects, AI employees, conversations, documents, tasks, and a plugin marketplace, letting users create projects, assign AI employees, collect business knowledge, start task conversations, and have AI employees carry out the actual work and deliver results. Its core goal is to help brick-and-mortar businesses drive AI transformation across IP building, video creation, public-domain customer acquisition, and private-domain conversion, cutting costs and boosting efficiency.

Process

Owned the full push from product positioning to runnable prototype: first pinned down how real-economy businesses use AI employees for IP building, video creation, public-domain acquisition, and private-domain conversion; then abstracted that into a product model of organizations, projects, AI employees, documents, and tasks, and set the boundaries for AI-employee capability isolation, cloud-persisted tasks, and traceable execution. On that basis, led multi-end delivery across desktop, website, admin console, cloud functions, and the WeChat mini-program, stringing organization management, project workspace, AI-employee collaboration, company knowledge base, task board, plugin marketplace, and public/private-domain operations into a business loop that actually runs.

Result

The project has closed the full loop from product definition and business model to multi-end delivery, with the desktop app, admin console, website, and mini-program all usable. It grounds organization and project management, AI-employee collaboration, task execution, knowledge, and public/private-domain operations in real-economy scenarios such as IP building, video creation, public-domain acquisition, and private-domain conversion. It now stands as a demo-ready, iterable AI-transformation workspace for real-world industries, laying the groundwork to help businesses cut costs and boost efficiency, reach real users, and validate the business.

Overview

OrderFi is Web3 store infrastructure for crypto-native merchants, cross-border sellers, and digital goods or service providers. It aims to let merchants create on-chain stores, list products, and accept stablecoin payments such as USDC with a Shopify-like experience, while reducing monthly platform fees, payment fees, settlement delays, KYC friction, and account-freezing risk. The system is built around wallet login, store contracts, product management, stablecoin checkout, order status, and instant settlement, giving merchants their own store contract and payment path.

Process

Owned the product positioning and MVP path definition: broke down merchant onboarding, contract deployment, product listing, customer checkout, USDC payment, gasless authorization, order management, and settlement flows; compared Shopify + Stripe cost structures with Web3 merchant pain points, and defined product hooks such as “store live in 3 minutes”, no monthly fees, and instant wallet settlement. Also shaped the Base L2 payment path, smart-contract store model, website demo, and testnet beta interaction narrative for validation.

Result

Formed an MVP framework from merchant onboarding to on-chain order fulfillment, supporting the website, demo materials, and Base testnet beta positioning. The work clarified low deployment cost, stablecoin settlement, deplatforming resistance, and future AI growth capabilities as differentiators, creating a product foundation for mainnet launch, merchant pilots, and business validation.

Overview

OpenSky Finance was an Aave-based NFT lending protocol for holders of blue-chip NFTs such as BAYC, MAYC, and CryptoPunks as well as liquidity providers. Its goal was to turn NFTs from assets that are mainly held or traded into on-chain financial assets that can be collateralized, borrowed against, and used for liquidity. The system combined peer-to-pool instant loans with peer-to-peer bespoke loans, allowing borrowers to unlock liquidity with NFT collateral while enabling lenders to participate in NFT-backed yield opportunities.

Process

Owned the product design and protocol-flow breakdown: defined core flows for NFT collateral, whitelisted assets, loan-to-value, interest, duration, pool-based borrowing, bespoke loan offers, repayment, liquidation, and defaults. Split the user journeys across borrowers, pool liquidity providers, and bespoke lenders, translating complex DeFi risk points into interaction flows and feature priorities that users could understand and engineering could execute.

Result

The protocol launched on Ethereum mainnet in August 2022, forming a full product loop from NFT collateral and credit evaluation to loan initiation, liquidity matching, and risk handling. After mainnet launch, TVL exceeded $30M, validating market demand for NFT-backed borrowing, pool liquidity, and risk-handling mechanisms. The project received grants from AAVE Grants DAO, Rarible Protocol, and Alchemy, and backing from MetaCartel Ventures, Alpha_nonce, FBG Capital, SNZ Holding, and other investors, with approximately $2M in disclosed funding.

Overview

Jike was an on-demand delivery fulfillment service for local services platforms, serving food delivery and local commerce scenarios such as Ele.me, Meituan Waimai, and Taodiandian. It aimed to solve last-mile delivery capacity shortages during order peaks, unstable city-level fulfillment, and the high cost of platform-owned delivery teams. The system was built around crowd-sourced couriers, order dispatch, fulfillment tracking, merchant collaboration, and city operations, enabling platform orders to be delivered through an external delivery network.

Process

As founder, owned zero-to-one product, operations, and commercialization: mapped the fulfillment chain across users, merchants, couriers, and platforms; built operating workflows for courier recruitment, order dispatch, fulfillment management, exception handling, and platform coordination. Also drove city expansion across Shanghai, Beijing, and Suzhou while handling business partnerships, customer communication, team organization, and fundraising.

Result

Built a crowd-sourced delivery fulfillment network that could be replicated across cities, scaling to 70,000+ registered couriers, 5,000 average DAU, and 20,000+ daily orders. Led the team to complete a multi-million RMB angel round from a well-known domestic investment institution in March 2015 and won a 300,000 RMB award at the Suzhou Industrial Park Yun Caifei startup competition.