AI Platform & Distributed Systems Engineer
我关注把模型、Agent、设备和分布式基础设施做成可运行、可观测、可恢复的系统。
- Robotics cloud platforms and device-facing distributed systems
- Model gateway、Agent runtime、memory、evaluation and cost control
- WebSocket/streaming workflows、observability and reliability engineering
Provider-neutral AI content pipeline with planning、queue、worker recovery、MCP、TTS、storage、rate limiting and cost boundaries.
Reproducible engineering labs for incident diagnosis、metrics、logs、token workflows and safe local experiments.
Self-hosted quota and cost dashboard for authorized AI providers，with Docker deployment and provider adapters.
XingP14 is my second GitHub account.
Projects maintained there are linked as second-account work, not presented as margrop originals.
XingP14/claude-code is not included in the selected work list or job-search portfolio.
- Portfolio: https://blog.margrop.net/career/
- Secrets stay outside repositories、logs and URLs.
- Experiments are labeled as experiments，not presented as production scale.
- Every automation path has an explicit permission boundary、failure mode and rollback path.