Incoming AI Graduate @ University of Adelaide | Ex-Baidu Intern
I am a software engineer passionate about bridging the gap between Cloud Infrastructure and Artificial Intelligence. With a background in developing high-availability cloud phone systems at Baidu, I am now focusing on building scalable AI applications.
cloud-ops-ai-agent — Async Industrial Execution Engine
An experimental framework merging Baidu Cloud-Phone operational patterns with modern async AI agent design. Built as a flagship project to demonstrate production-grade cloud orchestration:
| Feature | Implementation |
|---|---|
| Bounded Concurrency | asyncio.Semaphore for multi-device batch management |
| Three-Phase Safety Gateway | Risk classification → Identity verification → MFA (mirrors Baidu account security) |
| Cooperative Task Abort | CancellationToken + raise_if_cancelled() at high-frequency checkpoints |
| Config-Driven | Zero hardcoding — all params from config.json |
| Observability | TraceID full-chain, Prometheus metrics, AWS S3 log persistence |
| AWS Academy | S3 audit logs with Learner Lab session-token support |
Python · asyncio · boto3 · Prometheus · CI/CD (pytest, flake8)
- Languages: Python (asyncio, Deep Learning), TypeScript, Golang, SQL
- AI/ML: PyTorch, LangChain (RAG), OpenCV, Hugging Face
- Cloud/DevOps: AWS (EC2, S3, STS), Docker, Kubernetes, CI/CD (GitHub Actions)
- Front-end: React, Next.js, H5 Optimization (Performance Focused)
Role: Cloud Phone Product Optimization
- Scalability: Refactored static documentation into a cross-platform (PC/Android/H5) configurable back-end system, reducing deployment cycles.
- Security: Implemented advanced verification mechanisms for high-risk cloud phone operations, significantly enhancing user data safety.
- UX/DX: Designed task termination workflows for automation products, improving system controllability.
→ These patterns are now codified in cloud-ops-ai-agent as a reusable async execution engine.
