Mobius Gan

Context: Company in strategic transition from a single-protocol project toward an Agent OS + on-chain data infrastructure play. Sole PM responsible for defining and shipping 5 concurrent product lines. Key Deliverables: Cross-cutting Methodology Contributions:

Context: Early-stage Web3 on-chain reputation scoring project facing user cold-start + missing on-chain reputation data — a two-front challenge.

Context: Crypto retail traders facing high decision-making cost + steep barrier to off-chain data access.

Led advanced trading modules at CEX (one-click flip / trailing orders / batch take-profit & stop-loss / sub-accounts); planned multi-product flex-earning (dual-currency + structured); designed modular ops campaigns; led CCSwap DEX product design (lottery / leverage tokens / lending pools); SEO + content publishing on Next.js SSR.

Repositioned product from "data tool" to "on-chain intelligence layer for Agents"; shipped product site, Dashboard, and SDK documentation portal. Outcome: product became company's main revenue vehicle via Kafka data streams; led pricing benchmarking work that drove multiple enterprise leads into commercial negotiation.

Owned PRDs + interaction design for 6 core modules, all shipped — Swap, Perps, Prediction Markets, Smart Money Tracking, Referral, and Homepage V8 rebuild. Outcome: enables projects and KOLs to template-deploy professional Web3 sites in minutes; established as the company's Codeless platform core component library.

Defined product architecture from zero — conversation-page state machine, Skill system, configuration panel. Outcome: product in Beta, became core demo for fundraising materials.

Led Proof-to-Pay protocol product design; landed Solana + Base dual-chain escrow on mainnet with zero P0 incidents.

Designed 5-layer product architecture from scratch (Access / Data Intelligence / Risk / Execution / Settlement) with layered PRDs + 3 core Workflows (Alpha Discovery / Smart Money Copy / Portfolio Monitor). Outcome: serves as product baseline for company's flagship product over the next 12 months, now in development.

Codified a PM-led multi-source research approach — completed full Agent Wallet competitive analysis in half a day, output strategy MEMO that directly shaped company product direction. Defined 3-layer PRD methodology (Data Display / Interaction / Business Interface — each maintained independently) as the team's documentation standard, serving spec breakdown and delivery across all 5 product lines. Led ChainStream's data QA framework — designed 99 test cases across 12 validation dimensions for MCP / GraphQL / Kafka product interfaces, with AI-assisted automation for scripts and scheduled monitoring reports. Outcome: surfaced multiple financial-business data defects, established as the team's mandatory pre-delivery QA process for data products. AI Agent system (0→1): Defined Agent product form and user workflow from scratch — LLM selection (GPT-4o), Function Call architecture, prompt + parameter tuning, gamified onboarding with rewards + tutorials. Telegram Mini App (0→1): Owned product end-to-end from concept to launch — designed SBT-based on-chain reputation system, gamified task framework, partner-resource integrations. Outcome: 6K+ DAU / 180K+ cumulative users on a fixed-cost budget; reputation system became the company's core data asset, supporting subsequent fundraising and BD partnerships. LLM-driven AI trading products (0→1): Led design of dual product lines — AI Trading Assistant + Research Analyst; defined Agent off-chain data decomposition (social / news / macro) and multi-source communication architecture, with persona-driven personalization. Outcome: 500 DAU on a single channel / 5K+ cumulative tool users, validating AI Agent PMF in the crypto retail decision-making space. Token economics & commercialization design: Designed token supply thresholds, allocation schemes, sell-pressure assessment models, and airdrop plans; the token economic model provided the product foundation for the company's mainnet token launch, supporting commercialization strategy.

Designed for AI-Native independent writers — cut single-article production from 5-7h to ≤90min (4-5× single-creator efficiency). Owned full loop solo: user pain-point research → product architecture → spec definition → production deployment. Validated PMF for content-creator efficiency tools.

Designed for independent e-commerce / DTC sellers — cut single-video production from 5-8h to ≤60min (6-8× efficiency / 30-100× cost reduction). Tackled real pain points: slow trend response + high content production cost. Product covers trend scanning, product matching, video generation, and compliance review end-to-end.

Codified an original PM methodology (strategy decisioning / engineering breakdown) into executable protocols and local tools, forming a complete "methodology → tool → practice" loop that directly supports stable output of the two AI products above.

3 independent Python engines (polyarbbot / nbarbitrage / TSMOM) covering Polymarket binary options pricing, cross-source EV calculation, and time-series momentum factor research.