冯靖婉 AI 应用工程 / Agentic Engineer Coco Feng AI Applied Engineering · Agentic Engineer
设计师出身，扎根游戏行业，现以 AI 技术架构顾问身份为企业赋能。致力于将前沿技术栈转化为解决实际生产痛点的落地工具。
Designer by training, rooted in gaming, now building an AI education startup. Dedicated to translating cutting-edge tech into practical tools, while consulting as an AI Technology Architecture Advisor.
UI/UX 产品设计师
UI/UX Product Designer
MCP与API嵌入集成 · 本地模型
MCP & API Embeddings · Local Models
专注工具落地
Focusing on practical adoption
熟悉制作管线与痛点
Familiar with production pipelines
交互设计出身，游戏制作 × 产品设计 × AI 应用探索。在游戏公司积累项目管理与 AI 工程实践经验后，现以 AI 技术架构顾问身份为企业提供 AI 集成方案与技术咨询，同时创立觉然 JORAN 探索 AI 教育方向。
从 MCP 服务器开源项目到量化交易机器人，务实地将 AI 技术落地到各类实际场景中。
Interaction design background, bridging game production, product design, and AI applications. After gaining project management and AI engineering experience at a game studio, now serving as an AI Technology Architecture Consultant while founding JORAN to explore AI education.
From open-source MCP servers to quantitative trading bots — pragmatically shipping AI technology into diverse real-world scenarios.
- 自驱 AI 工程实践：在游戏制作职责外自学 AI 技能，实践 Agentic Workflow 与本地模型部署，善于借助 AI 工具检索 GitHub 开源资源并快速集成，产出多项面向实际生产场景的提效工具
- 百万字文本批量处理：使用 Google AI Studio 搭建基础工具流程，结合 Claude / Antigravity 对游戏剧本进行大批量文本校对与风险核查，交付内容经人工验证全部通过
- Agentic Workflow 实践：基于 Claude Code 构建多步骤 AI 任务流程，将内容生成与数据处理环节自动化
- AI 基础设施搭建：在 Hugging Face 选型并测试开源模型，使用 Ollama / LM Studio 在本地部署并运行推理，配置 Cloudflare 反向代理保障开发测试的网络访问
- 批量图像参考生成：为游戏 CG 绘图环节搭建批量参考图生成流程，大幅减少重复绘图时间，解决了生产中的实际瓶颈
- 项目管理：负责排期与资源分配，保障近期任务按时交付；协助管理外包供应商流程
- AI 技术架构顾问（2026.03 起）：转为顾问角色，为公司 AI 相关技术选型与集成方案提供指导；同期创立觉然 JORAN，探索 AI 教育方向
- Self-directed AI practice: Learned AI skills outside production duties, practicing Agentic Workflow and local model deployment. Regularly leverages AI tools to discover and integrate GitHub open-source resources, building practical productivity tools for real production scenarios
- Million-character text batch processing: Built a basic tooling pipeline using Google AI Studio, then used Claude / Antigravity to batch proofread and conduct risk checks on game scripts at scale — all deliverables passed manual verification
- Agentic Workflow in practice: Built multi-step AI task pipelines using Claude Code, automating content generation and data processing steps
- AI infrastructure setup: Evaluated and tested open-source models on Hugging Face, used Ollama / LM Studio to deploy and run local inference, and configured Cloudflare reverse proxy for stable dev/test network access
- Batch image reference generation: Built a batch reference image generation workflow for the game CG pipeline, significantly reducing redraw time and resolving a real production bottleneck
- Project management: Handled scheduling and resource allocation, ensuring on-time delivery of recent tasks; assisted in vendor management
- AI Tech Architecture Consultant (from Mar 2026): Transitioned to advisory role, providing guidance on AI technology selection and integration strategy; concurrently founded JORAN to explore AI education
- 参与《侏罗纪世界：进化3》（已发布）等国际游戏项目 VO/ADR 录音工作
- 严格执行数字资产管理规范，在国际化制作流程标准下保障交付质量
- VO/ADR recording for Jurassic World Evolution 3 (shipped) and other international game titles
- Strict digital asset management within an international game production pipeline
- 端到端设计执行：需求分析 → Figma 高保真 UI → 设计系统搭建
- 与产品经理和工程师协作，输出设计规范文档和部分 PRD
- End-to-end design ownership from requirements through high-fidelity Figma UI delivery; built and maintained design system
- Collaborated with PMs and engineers; produced design specs and contributed to PRD documentation
- 参与品牌焕新与官网重建，协助从策略到视觉执行的落地
- 统筹 Art Vancouver 2022 策展协助与现场执行，协调多方资源
- Contributed to brand refresh and website rebuild — supporting strategy through visual execution
- Coordinated Art Vancouver 2022: curation support, on-site execution, artist and logistics management
创立觉然 JORAN，面向非技术用户提供"从零开始把 AI 看明白"的课程体系。搭建 WeChat 数据解密 MCP 服务器、Cloudflare Pages 自动部署、AI 语音输入等工具链，用 Agentic Workflow 贯穿教学与工具开发。
Founded JORAN, offering "Understand AI from scratch" courses for non-technical users. Built WeChat decryption MCP server, Cloudflare auto-deployment, and AI voice input toolchain, powered by Agentic Workflow across teaching and development.
加密货币永续合约自动交易系统，多信号评分（EMA/RSI/布林带/波动率/情绪）驱动开仓决策，自适应止盈止损与追踪止盈机制，含实时仪表盘监控。
Automated crypto perpetual futures trading system with multi-signal scoring (EMA/RSI/Bollinger/volatility/sentiment), adaptive stop-loss/take-profit with trailing mechanism, and real-time monitoring dashboard.
调研并对比 ACE-Step 1.5 等开源音乐生成模型后，选定 MIDI 程序化生成路线。设计 Leitmotif 主导动机系统（统一音乐语言），通过 Python midiutil 批量生成 44 首 MIDI，经 FluidSynth + GeneralUser GS SF2 渲染为 MP3/WAV，作为项目制作参考交付。
After evaluating open-source models (ACE-Step 1.5, etc.), chose programmatic MIDI generation. Designed a Leitmotif system for thematic coherence, batch-generated 44 MIDI tracks via Python midiutil, rendered through FluidSynth + GeneralUser GS SF2 to MP3/WAV as production reference deliverables.
面向游戏 CG 绘图环节的美术参考批量生成工具。按角色特征批量输出多维度参考图，将手动搜图转化为自动化流程，大幅节约画师搜图时间。
Batch art reference generator for the game CG pipeline. Auto-generates multi-dimensional references by character traits, converting manual searching into an automated workflow.
搭建本地 AI 测试环境：对比测试开源模型，通过 Ollama / LM Studio 运行本地推理，配合 Cloudflare 反向代理解决接口访问限制。
Built a local AI test environment: benchmarked open-source models, ran local inference via Ollama / LM Studio, and configured Cloudflare reverse proxy for API access.
欢迎与我交流关于 AI 的机会 :D
Open to opportunities in AI.