Source: https://mayphus.org/profile/ Title: Profile Metadata: {"kind":"profile","language":"en","route":"/profile/","type":"page"} Mayphus Tang · 汤美法 Hangzhou, China Resume PDF (https://mayphus.org/profile/Mayphus_Meifa_Tang_CV.pdf) Email (mailto:mayphus.tang@gmail.com) GitHub (https://github.com/mayphus) One-page résumés by role: - Systems & infrastructure: [English](/profile/Mayphus_Tang_systems_infrastructure_EN.pdf) - DevOps & engineering platforms: [English](/profile/Mayphus_Tang_devops_engineering_platforms_EN.pdf) - Device & lab systems: [English](/profile/Mayphus_Tang_device_lab_systems_EN.pdf) Professional Summary Experienced across CI/CD platforms, production operations, Linux systems, networking, and embedded hardware. At State Street, supported an internal delivery platform spanning GitHub Enterprise, Jenkins shared libraries, Ansible, AWS and databases, tracing failures across build, scan, approval, release, and deployment stages. Earlier roles covered physical and virtual infrastructure, VPNs, containers, reverse proxies, databases, and board-level camera and recorder repair. Strongest at turning incomplete signals into a clear failing boundary, a safe recovery plan, and a reusable operational handoff. Core Skills Systems Linux, FreeBSD, TCP/IP, routing/firewalls, DNS/TLS, incident troubleshooting, ZFS CI/CD & Automation Jenkins, GitHub Enterprise, Ansible, YAML, Bash, Python, Racket, release troubleshooting Cloud & Platform AWS, EC2, Alibaba Cloud, Kubernetes, KVM, Docker/Podman Professional Experience - 2022 - 2025 DevOps / CI/CD Platform Engineer State Street Technology (Zhejiang) Co., Ltd. Technical environment: GitHub Enterprise, Jenkins shared libraries, YAML, Ansible, AWS, EC2, databases Databricks was also part of the team’s support scope. - Supported engineering teams using an internal CI/CD and self-service deployment platform built on GitHub Enterprise, Jenkins shared libraries, YAML, Ansible, and release approval gates. - Investigated failures across build, security scanning, approval, release, and AWS deployment stages, including EC2 and database workflows. - Supported CI/CD workflows and investigated failures, reading relevant implementation code to understand system behavior and guide troubleshooting. - Documented root cause, ownership boundary, safe workaround, reproduction path, and next action so teams could recover without relying on raw logs alone. CI/CD Failure Analysis Method - Start with the failed stage and its nearest authoritative signal: source or YAML input, shared-library execution, scan result, approval state, automation output, or downstream service response. - Compare expected inputs and outputs across adjacent stages to isolate the boundary between application configuration, delivery-platform behavior, automation, and dependent infrastructure. - Before recommending a retry, verify the current release state, the downstream resource condition, and whether the original failure left partial work that changes the recovery path. - Summarize each investigation as evidence, reproduction path, affected boundary, safe workaround, responsible owner, and next action rather than forwarding raw logs alone. - 2020 - 2022 Infrastructure Operations Engineer Hangzhou Fuzamei Tech Ltd. Technical environment: Linux, physical and virtual servers, Docker, Nginx, OpenVPN, WireGuard, monitoring, databases, blockchain - Delivered customer blockchain systems across cloud and on-premises environments, from infrastructure planning and cost estimates through deployment, node synchronization and troubleshooting. - Maintained company servers, KVM virtual machines, multi-office networks and employee workstations. - Managed routing, firewalls, OpenVPN, WireGuard, monitoring, Docker build environments, Nginx, databases, and the runbooks used for daily delivery. - Diagnosed deployments end to end across network, host, container, application, and data layers before production handoff. - Converted working configurations and verification commands into repeatable operating and recovery procedures. Infrastructure Deployment Verification Method - Model a deployment as one service path: domain and certificate, routing and firewall policy, VPN or public reachability, reverse proxy, host, container or process, application, and database. - Verify the path from the outside inward so a healthy local process is not mistaken for a reachable and production-ready service. - When a check fails, move to the adjacent layer and distinguish connectivity, host state, runtime execution, application behavior, and data dependency before changing configuration. - Use production-readiness checks to confirm routing, certificate validity, service configuration, database access, monitoring visibility, and the expected external response. - Capture the final working state, verification commands, rollback or recovery steps, and ownership notes in a reusable runbook. - 2018 - 2019 Technical Support / Camera, NVR & DVR Repair Engineer Zhejiang Dahua Vision Technology Co., Ltd. Technical environment: Linux, ARM, HiSilicon, u-boot, TCP/IP, FTP, cameras, NVR, DVR, embedded hardware - Diagnosed hardware and firmware faults in cameras, NVRs, DVRs, servers and network-connected embedded systems; repaired power circuits and replaced components, including BGA chips. - Used web-based firmware reflashing and UART/U-Boot recovery; recovered a camera by selectively reflashing firmware without removing its flash chip. - Validated full startup, storage, networking, recording and device functions before return to service. Embedded Device Diagnostic Method - Use a fixed diagnostic order: power and board condition, bootloader and boot progress, firmware, storage, network configuration, service connectivity, then device-specific functions. - Classify the first abnormal stage before replacing components, separating boot-chain, firmware, storage, configuration, hardware, and connectivity faults. - Use TCP/IP reachability and FTP transfer results as independent evidence so network or deployment-environment problems are not misclassified as device failure. - After a repair, repeat the full startup path and validate storage, networking, recording, and device-specific behavior rather than checking only the repaired component. - Apply the same sequence across cameras, recorders, servers, and other network-connected embedded devices while adapting the final functional checks to the equipment type. [My AI-assisted systems engineering learning plan](https://mayphus.org/ai-assisted-systems-engineering/) connects this earlier hands-on experience to the abilities I am developing now. Selected Public Work - Infrastructure (https://mayphus.org/infrastructure/) Designed and evolved my personal publishing infrastructure, experimenting with multiple publishing systems, frontend frameworks and hosting approaches before choosing Cloudflare. Publish the site and public APIs through Cloudflare Workers, with content, website deployment, and Input Foundry maintained separately. - Alpine Builder (https://mayphus.org/alpine-builder/) Built repeatable Alpine root filesystem and bootable image workflows for NanoPi boards. - Starflash (https://mayphus.org/starflash/) Created a safety-oriented macOS flashing tool with device validation, dry runs, and explicit write confirmation. - Input Method Builder (https://mayphus.org/typing/) Generated Rime and Yuanshu packages, keyboard previews, and downloads. Education Mechanical Design, Manufacturing and Automation 2014 - 2018 Changsha University - Built embedded and Linux systems through projects involving ROS robotics, 8051 MCU control and OpenWrt firmware. - Worked on Linux compatibility for a Huawei 4G module during university; able to build Linux kernels. - Combined mechanical, electronics, and software skills in engineering competitions and smart-agriculture automation projects. Languages & Credentials Chinese (native), English (CET-4), NCRE Level 2