雲端售後AI工程師（可實習）

半年以上在用 AI 寫程式的工具（Claude Code、Codex、OpenCode 之類）/ 6+ months vibing with AI coding tools (Claude Code, Codex, OpenCode, you name it)
1 年以上跟 AI 軟體混熟（ChatGPT、Claude、Kimi 都 OK）/ 1+ year hanging out with AI apps (ChatGPT, Claude, Kimi — all good)
資工 / 數學統計相關科系加分（不是硬門檻）/ CS / Math / Stats background is a bonus — not a hard gate

技術這塊 / Tech chops
Linux 基本功要有（檔案、行程、權限、網路會摸）/ Comfy with Linux basics (files, processes, permissions, networking — you get around)
以下會一樣就很讚 / Any of these and we're already smiling:

玩過 GPU / CUDA / NVIDIA 驅動 / Tinkered with GPU / CUDA / NVIDIA drivers
Docker 跟容器化部署 / Docker & containerized deploys
Python 看得懂、改得動 / Python — can read it, can break it, can fix it
碰過 PyTorch、Transformers / Have played with PyTorch & Transformers

軟實力 / Soft stuff that matters

會拆問題、找問題根源（這點超重要）/ Good at breaking problems apart and chasing down the root cause (this one's huge)
講話有條理，能把複雜的事講人話 / Can explain messy things in plain language
真心想幫使用者解決問題 / Actually cares about users
有一點產品 sense，看到怪怪的地方會想優化 / Has product instincts — spots rough edges and wants to smooth them

加分題 / Bonus points

玩過 GPU 雲 / 雲端運算 / 算力平台 / You've played around with GPU clouds / compute platforms
跑過 vLLM / Ollama / TGI 這類部署工具 / You've run vLLM / Ollama / TGI yourself
熟 RAG / Agent / 推理優化這類 AI 工作流 / Know your way around RAG / Agents / inference tuning
做過技術支援 / 客服 / 社群經營 / Done tech support / CS / community stuff before
看英文技術文件沒壓力 / English docs? No sweat

我們在找的人長這樣 / The vibe we're after

把 AI 當生產力外掛，不是拿來瞎聊而已 / Treats AI as a productivity weapon, not just a chat buddy
卡關時會自己動手查、動手試、動手驗證 / When stuck — searches, tries, verifies. Doesn't just wait for magic
對 AI / LLM / 算力基建真心有興趣，不是跟風 / Actually curious about AI / LLMs / compute infra — not just riding the hype

如果你每天就在用 AI 寫程式、調模型、折騰環境——那欸，這個位子根本為你準備的
If your daily routine is AI-assisted coding, model tweaking, and environment wrestling — yeah, this role's basically got your name on it.