Product-to-platform ownership
Aligning product, embedded R&D, vendors, factory systems, backend teams, and cloud architecture into one system that can actually ship.
Principal Cloud Platform Engineer · St. Gallen, Switzerland
I build backend and cloud systems that stay fast, reliable, and understandable.
Backend-focused engineer working across Go, AWS, Terraform, microservices, and distributed systems—from production architecture to the code and operational details that make it hold up.
Engineering focus
BuildGo · Node.js · Python
ScaleGlobal IoT · AWS · Distributed systems
OperatePerformance · Reliability · Security · Cost
15+ years turning difficult systems into useful products
Profile
A systems thinker who still likes shipping the code.
I started building and operating systems in my teens, then learned large-scale, low-latency engineering in Beijing. I work best at the seam between product, embedded devices, backend software, cloud infrastructure, and production operations.
I joined AirVisual/IQAir in Beijing as a Senior Software Architect in 2017. After moving to Switzerland as Cloud Architect, I designed and built the global IoT foundation for new device generations. From 2021, Lead DevOps was added to my architecture remit. Today I focus on hands-on backend and platform engineering as Principal Cloud Platform Engineer.
Where I create leverage
Engineering across the whole production path.
Backend & distributed systems
Building and operating Go, Node.js, and Python services with clear contracts, low latency, controlled failure modes, and safe migrations.
Production economics
Treating performance, reliability, scalability, security, and cost as one engineering problem—measured in live systems, not slide decks.
Experience
A career built from code to systems to global platforms.
IQAir is the latest chapter, not the beginning. Each role added a different layer: direct user responsibility, product engineering, scale, architecture, international alignment, and global platform ownership.
Hands-on principal engineering
Principal Cloud Platform Engineer
Focused on backend and cloud-native platform design, modernization, infrastructure, performance investigations, security, developer tooling, and production problem-solving across global services.
Lead DevOps mandate added from 2021
Cloud Architect
Owned architecture across cloud, backend, and IoT; developed technology direction with product and R&D; aligned international teams and vendors; advised on technical risk, security, reliability, cost, and production readiness; and mentored engineers while remaining hands-on.
Architecture, backend, and engineering practices
Senior Software Architect
Built and operated air-quality services, supported the move from on-premises infrastructure to AWS, and introduced code review, TDD, CI/CD, infrastructure as code, containerization, database indexing, and observability practices.
High-scale advertising, WiFi, and analytics systems
Senior Full Stack Developer
Where I learned scale directly: low-latency DSP/DMP/SSP services, APIs and captive portals used across major retail deployments, deployment automation and Nginx load balancing across hundreds of commodity servers, and technical leadership in a small team.
Commerce and operational software
C#.NET Main Developer
Built a localized e-commerce CMS, payment and CRM integrations, Windows Mobile barcode applications, database structures, security controls, documentation, and multi-region release processes.
Started building in my teens
Software, systems & hosting
Ran and extended an online-game hosting operation, then worked across Linux administration and hardening, PHP/.NET, SharePoint, ASP.NET, Android, MySQL, Apache, synchronization tooling, automation, and user support.
Selected work
Systems I can explain because I built and operated them.
Representative, non-confidential summaries. The emphasis is on the engineering decisions, ownership boundaries, and production outcomes behind the technology.
01 · Flagship system · Global IoT platform
One connectivity foundation from factory floor to worldwide production.
Designed and implemented the platform that allowed new IQAir device generations to bootstrap securely, connect to the closest regional broker, and integrate with the existing AirVisual backend ecosystem.
01Device & factory
Registration, provisioning, firmware identity
02Bootstrap service
Trust, configuration, nearest-region routing
03Regional IoT brokers
Active/active MQTT connectivity worldwide
04Backend ecosystem
Telemetry, control, data, and existing services
System decisions
- Defined device connectivity specifications, MQTT topic structure, telemetry contracts, bootstrap behavior, and regional routing.
- Designed active/active global connectivity while preserving compatibility with the existing device and backend platform.
- Created factory registration and provisioning flows, including the device key material accepted by the cloud trust chain.
End-to-end ownership
- Aligned external embedded and PCB vendors, internal R&D, product, factory, backend, and cloud stakeholders.
- Wrote the specifications and implemented core cloud, backend, and factory components rather than stopping at architecture.
- Carried the system into production with operability, security, reliability, scalability, and cost built into the design.
02
Backend performance & migration
Reducing a production footprint from 64 servers to 2
Rewrote a major device API from Node.js to Go and migrated it incrementally without breaking compatibility. In separate performance work, streamed database changes into Redis for a nearest-station service that moved lookups from seconds to microseconds.
03
Reliability, observability & FinOps
Making production behavior visible—and then affordable
Introduced infrastructure as code, CI/CD, tracing, structured logging, alerting, error tracking, service dashboards, autoscaling, and scheduled scaling. Built the first integrations and patterns, handled MongoDB and traffic incidents, and used operational evidence to improve reliability and cost.
04
Scale before cloud defaults
Learning distributed systems across hundreds of servers
At Goyoo, built high-scale advertising, WiFi, social, and analytics systems on commodity infrastructure. Developed a router API and captive-portal ecosystem used in KFC and Pizza Hut locations in China, with WeChat and Weibo login integrations.
05
Technical leadership
Connecting strategy to what reaches production
Worked across technology strategy, R&D direction, product and engineering alignment, architecture, technical risk, mentoring, vendor evaluation, code review, and production readiness—turning broad business goals into systems teams could implement and operate.
AI-assisted engineering
AI accelerates the loop. Structure keeps it reproducible.
I am developing a practical framework for building applications with AI agents: written specifications, milestone-based execution, self-contained prompts, and explicit verification instead of hidden conversational context.
Current tools: OpenAI Codex / GPT-5.6 Sol · ChatGPT · DeepSeek V4 Flash
- 01
Context as code
Capture constraints, specifications, decisions, interfaces, and acceptance criteria in versioned Markdown so the agent and the team share the same source of truth.
- 02
Milestone execution
Break work into bounded, reviewable milestones with self-contained prompts, explicit scope, and a verifiable definition of done instead of one oversized generation step.
- 03
Reproducible verification
Use repeatable builds, tests, review gates, and fresh-context checks so results can be reproduced, challenged, and handed to another engineer or agent without hidden context.
How I operate
Senior judgment, expressed in practical habits.
- 01
Solve the real problem
The work matters when it removes a visible constraint for a customer, teammate, or product—not when it merely adds technology.
- 02
Measure the system
Trace the real path, establish a baseline, and optimize the bottleneck—not the loudest theory.
- 03
Design for failure
Timeouts, retries, backpressure, idempotency, rollback, and degraded modes belong in the design from day one.
- 04
Own the outcome
Architecture, code, rollout, observability, security, and cost are parts of one production responsibility.
Technical range
Broad by experience. Deepest where systems meet production.
Technologies are grouped by how I use them, so earlier foundations and adjacent exposure are not presented as equal to current production depth.
Core today
Languages & runtimes
- Go
- Node.js
- Python
- TypeScript / JavaScript
- Bash
- SQL
Core today
APIs & distributed systems
- REST
- gRPC
- MQTT
- Microservices
- Event-driven systems
- API design
- Service templates
Deep production
Cloud platforms
- AWS
- ECS / Fargate
- Lambda
- CloudFront
- Route 53
- ALB / NLB
- VPC
- IAM
- AWS IoT
- S3
- SQS / SNS
- Kinesis
- WAF / Shield
- Azure / GCP exposure
Deep production
Infrastructure & delivery
- Terraform
- CloudFormation
- AWS CDK
- Serverless Framework
- Docker
- Bitbucket Pipelines
- CI/CD
- Git
- Trunk-based development
- Git Flow
Production
Data, search & messaging
- MongoDB Atlas
- DocumentDB
- DynamoDB
- PostgreSQL
- MySQL
- Cassandra
- Redis
- Elasticsearch / OpenSearch
- RabbitMQ
Production
Reliability & FinOps
- CloudWatch
- X-Ray
- Sentry
- Grafana
- OpenTelemetry
- Structured logging
- Alerting
- Autoscaling
- Cost allocation
- Incident analysis
Production
Security & identity
- Least privilege
- IAM Identity Center
- OIDC / OAuth 2.0
- Microsoft Entra ID
- WAF
- Shield
- Access patterns
- Linux hardening
Earlier foundations
Product & application stack
- C# / .NET
- WPF
- WCF
- ASP.NET MVC / WebAPI
- React
- PHP
- C / C++
- SharePoint
- Android
- Nginx
- Apache
Current practice
AI-assisted engineering
- OpenAI Codex
- GPT-5.6 Sol
- ChatGPT
- DeepSeek V4 Flash
- Markdown specifications
- Self-contained prompts
- Milestone plans
- Agent review gates
Best role fit
Senior Backend Engineer · GoStaff / Principal Platform EngineerCloud Infrastructure / SREIoT Platform Architect
Start a conversation