我的工作、演講和媒體報導My work, speaking engagements, and media coverage
Capture 是 Numbers Protocol 旗下的內容溯源工具，讓影像與數位媒體可以留下可查驗的來源脈絡。我曾擔任 Capture & AIKEA 群募專案行銷總監，負責把產品定位、上市節奏與市場溝通串起來。 Capture is Numbers Protocol's provenance tool for keeping verifiable context around images and digital media. I served as marketing director for the Capture & AIKEA crowdfunding projects, connecting product positioning, launch planning, and market communication.
Numbers Protocol 建構數位內容來源與驗證基礎建設，讓影像、創作與 AI 時代的內容可以保留來源紀錄。我負責全球成長、社群經營，以及把技術能力轉成不同產業能理解的敘事。 Numbers Protocol builds provenance and verification infrastructure for digital content, preserving source records for images, creative work, and AI-era media. I lead global growth, community, and the translation of technical capability into narratives different industries can use.
Omni Care 從團隊最卡住的一個流程開始，協助釐清需求、設計工作流、建置軟體或 AI 自動化，並把維護責任說清楚。我參與把 Numbers 內部 AI 導入與工程治理經驗，整理成外部團隊也能採用的合作方式。 Omni Care starts with one stuck workflow, then helps teams clarify needs, design operations, build software or AI automation, and define maintenance ownership. I help turn Numbers' internal AI adoption and engineering-governance practice into a collaboration model other teams can use.
Creative Origin Alliance 連結創作者、技術團隊與平台，討論創作內容在 AI 訓練、生成與授權環境中的來源與權利問題。我曾任 Creative Origin 行銷創意總監，也是 Creative Origin Alliance 創始成員。 Creative Origin Alliance connects creators, technologists, and platforms around provenance and rights questions in AI training, generation, and licensing. I served as Marketing Creative Director at Creative Origin and am a founding member of the Creative Origin Alliance.
DT42 是我早期參與的 AI 新創，聚焦智慧邊緣運算與硬體場景。我參與業務開發與策略成長，協助團隊把技術能力整理成市場可以理解的定位與故事。 DT42 was an early AI startup I worked with, focused on smart edge computing and hardware scenarios. I worked on business development and strategic growth, helping turn technical capability into market positioning and narrative.
AIKEA 是一個 Kickstarter 群募專案，嘗試把主打隱私的 AI 相機帶進家庭場景。我以 Capture & AIKEA 群募專案行銷總監身分負責產品上市與行銷活動，協助把技術產品轉成一般消費者能理解的使用情境。 AIKEA was a Kickstarter campaign exploring privacy-focused AI cameras for home scenarios. As marketing director for the Capture & AIKEA crowdfunding projects, I led launch and campaign work that translated a technical product into consumer use cases.
共同創辦的 YouTube 頻道，用短影片把 AI 概念、工具觀察和使用情境講成比較容易進入的版本。 Co-created a YouTube channel that explains AI concepts, tool observations, and use cases through short, accessible videos.
和團隊一起看客服、營運、內容生產與知識管理流程，先找出痛點、資料條件和 owner，再決定哪個場景值得優先導入 AI。Review customer-service, operations, content-production, and knowledge-management workflows with the team, then identify pain points, data conditions, owners, and practical AI starting points.
把 agent 的角色、權限、交接、稽核紀錄與人類 review 點寫清楚，讓 AI 工作不是一次性 demo，而是能被日常接住的流程。Define agent roles, permissions, handoffs, audit trails, and human review points so AI work is not a one-off demo, but a workflow daily operations can absorb.
協助團隊把 AI 解決方案放回自己的資料、流程、權限與預算條件裡比較，避免只看 demo 效果就做採購決策。Help teams compare AI solutions against their own data, workflows, permission model, and budget conditions, instead of buying from the demo alone.
協助新創與產品團隊判斷哪些情境適合用 Vibe Coding 快速試錯，哪些只要牽涉金流、資安、資料品質或長期維護，就必須設下工程檢核點。Help startup and product teams decide where vibe coding is useful for rapid experiments, and where payments, security, data quality, or long-term maintenance require explicit engineering review points.
把 Transparent、Auditable、Explainable、Agentic 四個原則，轉成角色分工、紀錄格式、升級規則與定期 review 節奏。Turn Transparent, Auditable, Explainable, and Agentic principles into role design, record formats, escalation rules, and regular review cadence.
帶非工程團隊用 AI 做出小型可運作原型，同時練習怎麼描述需求、檢查輸出、判斷風險與交接給工程或營運 owner。Guide non-engineering teams to build small working prototypes with AI while practicing requirements, output review, risk judgment, and handoff to engineering or operations owners.
POLITICO
討論中國針對台灣選舉的假訊息攻勢 On China's disinformation campaign targeting Taiwan's elections
AmCham
討論打擊選舉假訊息和數位驗證的重要性 On fighting election disinformation and why digital verification matters
Meet Global
分享區塊鏈技術和數位媒體驗證的見解 Insights on blockchain and digital media verification
田間 Tian Jian
為何要保存資料檔案？ Why preserve data archives?
Meet 創業小聚
嚴世紀觀點：當海水退潮後，NFT 是泡沫還是創新？ Sofia Yan's view: when the tide goes out, is NFT a bubble or an innovation?
MakerPRO
【OpenVINO】Edge AI 應用落地與開發實務論壇報導（DT42 時期） [OpenVINO] Edge AI deployment & development forum report (DT42 era)
TEC TALK 創業不歸路
敏捷開發思維落實遠距制度設計：專訪數據溯源新創「Numbers」共同創辦人 Sofia Yan Agile thinking for remote-first design: an interview with Numbers co-founder Sofia Yan
區塊勢
以代幣經濟為數位影像加上溯源履歷 ft. Numbers Protocol 共同創辦人 Sofia Yan Adding provenance to digital images with token economics, ft. Numbers co-founder Sofia Yan
數位關鍵字
文組人也能上手的 Vibe Coding！AI 如何協助行銷人員解決重複性工作 Vibe coding even humanities folks can do: how AI helps marketers with repetitive work
數位關鍵字
圖、詐騙、假新聞越來越多！技術能重建你我對網路資訊的信任？ Deepfakes, scams, and fake news: can technology rebuild our trust in online information?
職場來一課
文組生也能創建自動化流程工具！AI 如何融入工作流，提高做事效率 Even humanities grads can build automation tools: weaving AI into workflows for efficiency
de lab cafe
de lab 輕鬆聊：科技麻瓜在數據新創的生存指南 de lab chat: a tech muggle's survival guide inside a data startup
Numbers Network
新世代新聞驗證：科技如何重塑我們對真實的認知 Next-gen news verification: how technology reshapes our sense of what's real
創業新聲帶
Numbers (上) │ 如何擊破假新聞？Numbers 推出「真相快門」以區塊鏈技術確保數據完整性 Numbers (Part 1) │ Breaking fake news: the blockchain "truth shutter" for data integrity
創業新聲帶
Numbers (下) │ 如何擊破假新聞？Numbers 推出「真相快門」以區塊鏈技術確保數據完整性 Numbers (Part 2) │ Breaking fake news: the blockchain "truth shutter" for data integrity
SpeakingSpeaking
演講與工作坊會把這些 AI 導入、內容溯源、團隊協作的現場，整理成團隊聽得懂、也能帶回去試的版本。 Talks and workshops turn these AI adoption, content provenance, and team-operation field notes into something teams can understand and try.