隽戈

隽戈

Cloud-Native Architect / SRE Engineer / Tech Blogger
A hybrid profile combining SRE, Cloud-Native, Observability, Data Engineering, Machine Learning and Platform Development. Long-term focus on large-scale distributed system stability governance and AIOps platform construction, with strong platform engineering capabilities to independently complete AIOps system architecture design, core module development, and stability assurance.

Core Capabilities

🛡️ SRE Stability Governance ☁️ Cloud-Native Architecture
📊 Observability Platform 📐 Data Engineering
🤖 Machine Learning 🔧 Platform Development

Tech Stack

Languages
Python Golang Java JavaScript Terraform Ansible
Core Domains
Kubernetes AIOps Prometheus OpenTelemetry Grafana Platform Engineering
Observability
Metric Log Trace Event ELK Skywalking
Data & Messaging
Kafka Flink ClickHouse Elasticsearch
CI/CD & DevOps
Jenkins GitLab CI ArgoCD Helm Kustomize

Community & Open Source

Active Communities
  • Expert Ops Community
  • HAMI Project Community
  • Xinference Community
  • CNCF Cloud-Native Southwest China
  • ArkSphere AI Community
Open Source Experience
kubeasz Project-HAMi higress vllm sglang xinference OpenKruise
Hi, I'm JunGe.

My Engineering Philosophy

In the AI era, a developer's role shifts from writing code to making strategic design decisions. AI is your tactical executor—it writes code, runs tests, and refactors—but strategic-level decisions must be made by humans. The following four principles are the underlying logic behind my engineering decisions:

🔷 I Design, Not Pile Up

A Deep Module—rich in functionality but simple in interface—beats a sea of shallow, fragmented code. Encapsulate complexity behind the interface, so each interaction only requires understanding the interface, not every detail.

🔷 I Unify, Not Guess

People, code, and AI speak the same vocabulary. A Ubiquitous Language isn't documentation—it's the engineering discipline of eliminating ambiguity. In the AI era, this value is amplified tenfold.

🔷 I Think First, Then Act

Let the design be interrogated before writing code. Dependencies, edge cases, data models— AI won't make decisions for you; it only accelerates the decisions you've already made. If you haven't thought it through, AI only accelerates the chaos.

🔷 I Verify, Then Deliver

Every step is verifiable. In an era where AI generates hundreds of lines of code, TDD's role shifts from quality assurance to process control—keeping every step within control and correcting deviations the moment they appear.

These principles aren't new inventions—they come from Domain-Driven Design, A Philosophy of Software Design, and Extreme Programming. In the AI era, they haven't been made obsolete—they've become more important.

Career Background

Held a key technical leadership role in Singapore Telecom Smart City Project, driving the complete infrastructure evolution from early Mesos to modern Kubernetes cloud-native architecture.

During tenure at Ant Group and leading internet banks, deeply involved in cloud-native transformation of financial-grade core systems. Focused on high-availability infrastructure and platform engineering capabilities, significantly improving delivery efficiency and system elasticity while maintaining financial-grade stability.

Featured Projects

Enterprise DevOps Platform
Architecture / Implementation / DevOps Level 3 Certification

Built enterprise-class private R&D workflow platform from scratch based on GitOps + CI/CD + Kubernetes. Standardized pipelines and automated delivery systems significantly shortened development cycles. Led platform to achieve DevOps Level 3 Certification, establishing industry-leading engineering standards.

Enterprise DevOps Platform architecture diagram
AI Heterogeneous Computing Infrastructure
HAMi / vLLM / sglang / GPU Scheduling

Built high-performance heterogeneous computing platform based on HAMi + Kubernetes, achieving vGPU resource pooling and dynamic elastic scheduling. Successfully deployed vLLM / sglang LLM containerization solutions, providing unified, efficient, and scalable computing infrastructure for multi-scenario inference tasks.

AI Heterogeneous Computing Infrastructure architecture diagram
Next-Gen AIOps System
Dify / OpenClaw / Harness / n8n / LLM Application

Explored deep LLM applications in operations domain using Dify / OpenClaw / Harness / n8n framework. Implemented intelligent log attribution analysis, automated fault diagnosis, and self-service ops chatbot, building an AIOps closed-loop system from reactive response to proactive governance.

Next-Gen AIOps System architecture diagram
Data Center Equipment Management System (Vibe Coding)
Codex / Claude Code / Windsurf / Redfish / DCIM

Using Vibe Coding (AI-assisted development with Codex / Claude Code / Windsurf), built a Data Center Infrastructure Management (DCIM) system from scratch. It covers full-category asset management—unified multi-brand management of rack equipment, security appliances, storage devices, and network devices; integrates out-of-band BMC data via the standard Redfish protocol and automated crawlers for automated collection and ops workflow optimization; and provides data-center rack topology visualization, combined with MAC address-based automatic network link tracing, significantly improving asset visibility and operational efficiency.

Personal Vision

Beyond work, committed to connecting with the community through video content creation and technical sharing. Continuously producing high-quality tech Vlogs, sharing practical experience in a visual and systematic way to promote cloud-native and AI technology adoption.

  • 📚 Knowledge: Building systematic tech knowledge base
  • 💡 Insights: Sharing industry trend perspectives
  • 📷 Life: Documenting colorful moments beyond technology