Senior engineer (10+ yrs) — built a distributed platform from scratch, contributed to OpenStack, now extending into AI platform & agent engineering.
资深工程师(10 年+)— 从零造过分布式平台、OpenStack 开源贡献者,正把这套能力延伸到 AI 平台与 Agent 工程。
I've spent the last decade across the full stack: Python backends, platform and distributed infrastructure, and DevOps automation — taking large projects from architecture through delivery into production, and leading the teams (5+) behind them. These days I'm pointing that foundation at AI: building agents and the infrastructure around them, as the natural next layer on top of the platform and distributed-systems work I already know well.
我的十年覆盖了全栈:Python 后端、平台与分布式基础设施、DevOps 自动化——把大中型项目从架构、开发一路带到上线,并组建带过团队(5+)。现在我把这套底子对准 AI:做 Agent 和它周边的基础设施,作为平台与分布式经验之上自然的下一层。
-
Platform & distributed systems — Designed and built a distributed compute platform end to end: a custom Docker scheduler (scheduling, metering, reporting, APIs), Ceph distributed storage, AMQP messaging, and a Celery batch-processing subsystem behind an LB/HA cluster.
平台与分布式系统 — 端到端设计并实现了一套分布式计算平台:自研 Docker 调度器(调度 / 计费 / 报表 / API)、Ceph 分布式存储、AMQP 消息、Celery 批处理子系统,跑在 LB/HA 集群之上。 -
DevOps & infrastructure automation — OpenStack, Docker/Kubernetes/Swarm, and one-click automated deployment with Cobbler/Puppet/Ansible/Jenkins — from bare metal and IDC operations to cloud, reliability and automation at scale.
DevOps 与基础设施自动化 — OpenStack、Docker/Kubernetes/Swarm,以及基于 Cobbler/Puppet/Ansible/Jenkins 的一键自动化部署——从物理机、IDC 运维到云,把规模化的稳定性与自动化做扎实。 -
Full-stack delivery — Python (Django/Flask) backends with Vue/AngularJS frontends, shipped end to end — owning a system from requirements and architecture to the UI users actually touch.
全栈交付 — Python(Django/Flask)后端 + Vue/AngularJS 前端,端到端落地——从需求、架构一直管到用户真正触达的界面。
-
Self-built distributed platform (FUVISM) — Architected and led a Docker-based "immersive" online-education platform, then built and ran the 9-person team behind it. Custom scheduler + Ceph + AMQP + Celery, with the user-facing system isolated from the compute backend over internal REST.
自研分布式平台(FUVISM) — 主导了一套基于 Docker 的"沉浸式"在线教育平台的架构与开发,并组建带领 9 人团队。自研调度器 + Ceph + AMQP + Celery,业务系统与计算后端通过内部 REST 隔离。 -
OpenStack contributor — 58+ commits to the OpenStack community; optimized Horizon API integration and built automated deployment tooling. Later worked at the OpenAtom Foundation on open-source infrastructure and ecosystem projects.
OpenStack 开源贡献者 — 向 OpenStack 社区贡献 58+ commits,优化 Horizon API 整合、构建自动化部署工具;后在开放原子开源基金会做开源基础设施与生态项目。 -
CMDB asset-management system (Megvii) — Owned an asset platform managing 5,000+ devices and 500M+ RMB in assets, covering the full asset lifecycle — design, agents for system/network/hardware collection, and operations.
CMDB 资产管理系统(旷视) — 负责管理 5,000+ 设备、5 亿+ RMB 资产的平台,覆盖资产全生命周期——设计、系统/网络/硬件信息采集 Agent、运维保障。 -
Platform engineering @ HTX (Huobi) — Building and maintaining internal engineering-efficiency, ticketing, and release platforms. Working inside the crypto industry, with a working knowledge of BTC/Ethereum/Hyperledger consensus and Solidity smart contracts.
平台开发 @ HTX(火币) — 负责公司内部工程效能、工单、发布平台的开发与维护;身处加密行业,了解 BTC/Ethereum/Hyperledger 共识与 Solidity 智能合约。
A hand-rolled AI agent in Python (DeepSeek) — already does tool calling, multi-step reasoning loops, and RAG. Growing it from a clean from-scratch core toward something engineered and deployable: a small, readable, runnable open-source agent. (Work in progress.)
正在手搓一个 Python + DeepSeek 的 AI Agent —— 已实现工具调用、多步推理循环和 RAG。按"手搓内核 → 工程化 → 可部署"推进,目标做成一个干净、可跑、可讲的开源 agent。(进行中)
Languages
Backend
Platform & Infra
DevOps
Frontend
