[{"data":1,"prerenderedAt":341},["ShallowReactive",2],{"site-info":3,"work-embedded-vision-background":99},{"id":4,"company":5,"extension":16,"meta":17,"services":18,"social":96,"stem":97,"__hash__":98},"site\u002Fsite.yml",{"name":6,"nameEn":7,"email":8,"address":9,"addressEn":10,"lineUrl":11,"lineId":12,"duns":13,"youtube":14,"youtubeId":15},"銡策數位有限公司","Jecision Digital Co., Ltd.","contact@jecision.com","231 新北市新店區北新路二段 262 號 3 樓之 3","3F.-3, No. 262, Sec. 2, Beixin Rd., Xindian Dist., New Taipei City 231, Taiwan","https:\u002F\u002Fline.me\u002FR\u002Fti\u002Fp\u002F@374qnumi","@374qnumi","963905705","https:\u002F\u002Fwww.youtube.com\u002Fshorts\u002F6UE0WTmyjsg","6UE0WTmyjsg","yml",{},[19,43,69],{"key":20,"zh":21,"en":32},"audit",{"title":22,"tagline":23,"summary":24,"points":25,"deliverable":29,"detail":30,"duration":31},"企業 AI 導入健檢","先看清楚，再花錢","60 分鐘線上訪談，盤點最耗時、最依賴人力的營運環節；我們會告訴你哪些工作現在就能交給 AI、哪些還不成熟。",[26,27,28],"現況盤點：梳理日常營運中最耗時、最依賴人力的環節","痛點診斷：判斷哪些工作現在就能交給 AI，哪些建議先保留人力","可行性評估：依你的業務特性與預算，篩選適合的模型與工具","《企業 AI 導入建議報告書》（PDF）","含 1–2 個「開發成本最低、成效最明顯」的優先項目、技術棧建議、預期節省的人力時間與下一步","訪談 60 分鐘，報告書 3 個工作天內交付",{"title":33,"tagline":34,"summary":35,"points":36,"deliverable":40,"detail":41,"duration":42},"AI Readiness Audit","Look before you spend","A 60-minute remote audit of the workflows that eat the most time and people-power — and an honest answer on what AI can take over today, and what it cannot.",[37,38,39],"Audit the daily workflows that consume the most time and headcount","Triage which tasks AI can own today, and which should stay human","Shortlist the models and tools that fit your budget and constraints","AI Adoption Recommendation Report (PDF)","One or two highest-ROI entry points, a recommended tech stack, expected hours saved, and the concrete next step","60-minute call, report delivered within 3 business days",{"key":44,"zh":45,"en":56},"prompt",{"title":46,"tagline":47,"summary":48,"points":49,"deliverable":53,"detail":54,"duration":55},"提示詞工程","讓 AI 穩定產出，而不是碰運氣","針對你的實際任務設計提示詞與多步驟工作流，並做極限測試，確保換了資料、換了人操作也不會走鐘。",[50,51,52],"客製提示詞：ChatGPT、Claude、Gemini、Midjourney 等主流工具的專屬模板","除錯與優化：修掉幻覺、格式跑掉、時好時壞的問題","工作流建置：把複雜任務拆成多步驟提示詞與可重複的 SOP","提示詞文件或設定檔","附使用說明與範例，並提供 1 對 1 教學，確保你與團隊能自己維護","依任務複雜度報價，通常 3–7 個工作天",{"title":57,"tagline":58,"summary":59,"points":60,"deliverable":66,"detail":67,"duration":68},"Prompt Engineering","Make AI produce reliably, not by luck","Prompt sets and multi-step workflows designed for your actual tasks, stress-tested so they keep working when the data or the operator changes.",[61,62,64],"Custom prompts for ChatGPT, Claude, Gemini, Midjourney and friends",{"Debugging and tuning":63},"kill hallucinations, broken formats and flaky output",{"Workflow design":65},"break complex work into repeatable multi-step SOPs","Prompt library or exportable config","Ships with usage notes and examples, plus a 1:1 walkthrough so your team can maintain it","Quoted by complexity, typically 3–7 business days",{"key":70,"zh":71,"en":82},"poc",{"title":72,"tagline":73,"summary":74,"points":75,"deliverable":79,"detail":80,"duration":81},"極速 PoC","五天，看到能跑的東西","不談架構藍圖，直接做一個能動的雛形。五個工作天內交付線上可操作的智能體，用你真實的資料驗證。",[76,77,78],"Day 1：15 分鐘對焦，鎖定單一痛點與資料盤點","Day 2–4：模型串接、邏輯設計與雛形開發","Day 5：Demo 展示與 AI 擴展建議書","可線上操作的智能體雛形","交付範圍為概念驗證與展示，不含後續維運或大規模伺服器部署；要進階成完整系統會另行報價","5 個工作天",{"title":83,"tagline":84,"summary":85,"points":86,"deliverable":93,"detail":94,"duration":95},"Rapid PoC","Five days to something that runs","No architecture decks — a working prototype built in five business days, validated against your real data.",[87,89,91],{"Day 1":88},"15-minute scoping call, one pain point, data inventory",{"Day 2–4":90},"model integration, logic design, prototype build",{"Day 5":92},"live demo plus an AI expansion recommendation","A runnable online agent prototype","Proof of concept and demo scope only; ongoing operations and large-scale deployment are quoted separately","5 business days",[],"site","XTwARfVAOcrXNiKehsM6V6cdkd2HXy4d5yZVrSWrI3Q",{"zh":100,"en":226},{"id":101,"title":102,"body":103,"client":197,"cover":198,"coverAlt":199,"description":200,"draft":201,"extension":202,"industry":203,"lang":204,"meta":205,"navigation":206,"order":207,"path":208,"results":209,"seo":219,"services":220,"slug":222,"stem":223,"translationKey":224,"year":224,"__hash__":225},"work\u002Fwork\u002Fembedded-vision-background.md","開公司之前：二十年嵌入式與影像辨識的底子",{"type":104,"value":105,"toc":188},"minimark",[106,110,119,122,125,141,148,151,157,163,169,172,179,182,185],[107,108,109],"h2",{"id":109},"為什麼要寫這一頁",[111,112,113,114,118],"p",{},"很多人以為「AI 顧問」的工作是選模型、寫 prompt。但真正讓專案失敗的，通常不是模型不夠強，而是",[115,116,117],"strong",{},"沒人處理模型以外的那 90%","：裝置跑不動、光線一變就失效、資料格式對不上、上線後沒有人會維護。",[111,120,121],{},"我們的底子是在那裡練出來的。",[107,123,124],{"id":124},"當時的條件",[126,127,128,132,135,138],"ul",{},[129,130,131],"li",{},"沒有 GPU，推論要擠進 CPU 的預算裡",[129,133,134],{},"記憶體以百 MB 計，模型不能「先載入再說」",[129,136,137],{},"裝置在現場：會震、會熱、會斷電",[129,139,140],{},"沒有人會在你旁邊看 log，錯了就是產線停",[111,142,143,144,147],{},"在這種環境裡，「準確率 95%」不是終點，",[115,145,146],{},"「連續跑三個月不出事」才是","。",[107,149,150],{"id":150},"這段經歷留下來的三個習慣",[111,152,153,156],{},[115,154,155],{},"一、先量測，再優化。"," 在邊緣裝置上，憑感覺優化通常會優化錯地方。先找出真正的瓶頸——是前處理、是模型、還是 I\u002FO。",[111,158,159,162],{},[115,160,161],{},"二、為失敗設計。"," 現場沒有人會看 exception。模型信心不足時要能退回一個安全的行為，而不是整個流程卡住。",[111,164,165,168],{},[115,166,167],{},"三、交付要能被接手。"," 代碼、參數、資料格式、怎麼重新訓練——如果對方接不了，這個專案就沒有結束的一天。",[107,170,171],{"id":171},"和現在的工作有什麼關係",[111,173,174,175,178],{},"現在做 LLM 專案，面對的其實是同一組問題：模型只是其中一塊，",[115,176,177],{},"資料從哪來、錯的時候怎麼辦、誰來維護","才是決勝點。",[111,180,181],{},"這也是我們堅持「先健檢、再 PoC」的原因——先確認問題是真的、資料是有的、人願意改流程，再去談模型。",[107,183,184],{"id":184},"目前沒做的",[111,186,187],{},"這一頁還沒有放具體的專案細節（客戶、裝置、量測數字）。那些需要取得客戶同意才能公開；如果你們手上也有類似條件的專案，歡迎直接聊。",{"title":189,"searchDepth":190,"depth":190,"links":191},"",2,[192,193,194,195,196],{"id":109,"depth":190,"text":109},{"id":124,"depth":190,"text":124},{"id":150,"depth":190,"text":150},{"id":171,"depth":190,"text":171},{"id":184,"depth":190,"text":184},"團隊背景","\u002Fimages\u002Fwork\u002Fembedded-vision-background.svg","嵌入式影像辨識示意圖","在成立銡策之前，團隊做的是資源受限裝置上的即時影像辨識——沒有 GPU、記憶體只有幾百 MB，還要在真實的光線與震動下跑得穩。這段經歷決定了我們後來做 AI 專案的方式。",false,"md","嵌入式系統","zh-TW",{},true,3,"\u002Fwork\u002Fembedded-vision-background",[210,213,216],{"label":211,"value":212},"累積時間","20+ 年",{"label":214,"value":215},"主要場景","資源受限裝置",{"label":217,"value":218},"擅長","把模型調到能上線",{"title":102,"description":200},[221],"edge","embedded-vision-background","work\u002Fembedded-vision-background",null,"fS1oozwrTs7BE73bTi2eqs5QigRLHVpHU5ZmOvOodR8",{"id":227,"title":228,"body":229,"client":319,"cover":320,"coverAlt":321,"description":322,"draft":201,"extension":202,"industry":323,"lang":324,"meta":325,"navigation":206,"order":207,"path":326,"results":327,"seo":337,"services":338,"slug":222,"stem":339,"translationKey":224,"year":224,"__hash__":340},"work\u002Fwork\u002Fen\u002Fembedded-vision-background.md","Before the company — twenty years of embedded vision",{"type":104,"value":230,"toc":312},[231,235,242,245,249,263,269,273,279,285,291,295,302,305,309],[107,232,234],{"id":233},"why-this-page-exists","Why this page exists",[111,236,237,238,241],{},"People assume an \"AI consultant\" is someone who picks models and writes prompts. But projects rarely fail because the model is not strong enough. They fail because ",[115,239,240],{},"nobody handled the other 90%",": the device cannot run it, the lighting changes and it breaks, the data format does not match, and after go-live nobody maintains it.",[111,243,244],{},"That is where our instincts came from.",[107,246,248],{"id":247},"the-conditions-we-worked-in","The conditions we worked in",[126,250,251,254,257,260],{},[129,252,253],{},"No GPU — inference had to fit inside a CPU budget",[129,255,256],{},"Memory measured in hundreds of megabytes; \"just load the model\" was not an option",[129,258,259],{},"Devices in the field: vibrating, hot, occasionally losing power",[129,261,262],{},"Nobody watching the logs; when it is wrong, the line stops",[111,264,265,266],{},"In that environment, \"95% accuracy\" is not the finish line. ",[115,267,268],{},"Running for three months without an incident is.",[107,270,272],{"id":271},"three-habits-it-left-behind","Three habits it left behind",[111,274,275,278],{},[115,276,277],{},"Measure before you optimise."," On an edge device, optimising by intuition usually optimises the wrong thing. Find the real bottleneck first — pre-processing, the model, or I\u002FO.",[111,280,281,284],{},[115,282,283],{},"Design for failure."," Nobody on site reads exceptions. When confidence is low, the system needs a safe fallback, not a stalled pipeline.",[111,286,287,290],{},[115,288,289],{},"Deliver something that can be taken over."," Code, parameters, data formats, how to retrain. If the client cannot maintain it, the project never really ended.",[107,292,294],{"id":293},"how-it-connects-to-the-work-we-do-now","How it connects to the work we do now",[111,296,297,298,301],{},"LLM projects face the same shape of problem: the model is one piece. ",[115,299,300],{},"Where the data comes from, what happens when it is wrong, and who maintains it"," are what decide the outcome.",[111,303,304],{},"That is why we insist on audit-before-PoC: confirm the problem is real, the data exists, and someone is willing to change the process — before talking about models.",[107,306,308],{"id":307},"what-is-not-done","What is not done",[111,310,311],{},"This page does not yet include specific project details (clients, hardware, measured numbers). Those need client consent to publish. If you have a project with similar constraints, let's talk.",{"title":189,"searchDepth":190,"depth":190,"links":313},[314,315,316,317,318],{"id":233,"depth":190,"text":234},{"id":247,"depth":190,"text":248},{"id":271,"depth":190,"text":272},{"id":293,"depth":190,"text":294},{"id":307,"depth":190,"text":308},"Team background","\u002Fimages\u002Fwork\u002Fembedded-vision-background-en.svg","Embedded vision on constrained hardware","Before Jecision existed, the team built real-time vision on constrained devices — no GPU, a few hundred MB of RAM, and real light and vibration. That background shapes how we run AI projects today.","Embedded systems","en",{},"\u002Fwork\u002Fen\u002Fembedded-vision-background",[328,331,334],{"label":329,"value":330},"Experience","20+ years",{"label":332,"value":333},"Typical setting","Constrained devices",{"label":335,"value":336},"Strength","Getting models into production",{"title":228,"description":322},[221],"work\u002Fen\u002Fembedded-vision-background","xa8RvuMExuycOr-qLpI4bUHzUxNo3uTEqUoigg8FYlw",1792775653389]