CURRENT PRACTICE
Python backend systems
Typed services, data lifecycles, storage boundaries, and failure-aware workflows.
PYTHON SYSTEMS ENGINEER / AI INFRASTRUCTURE
Ray Carter
Building systems around AI with receipts.
I build typed Python services, resilient data ingestion, and agent workflows around explicit contracts, tests, and failure modes.
PROFILE
DIRECTION
AI infrastructure
Applying backend depth to the systems that make AI products dependable.
BACKGROUND
MSc, Data Analytics & AI
Based in Hong Kong · Working in Chinese and English.
AT A GLANCE
- 03
- Selected, documented systems
- v3.0.0
- ExcelAlchemy published release
- UTC+8
- Hong Kong base
- 02
- Working languages
SELECTED WORK
BACKEND / INFRASTRUCTURE
Upload Control Plane (opens in a new tab)
A Python-first multipart upload system with resumable clients, explicit lifecycle state, and failure tests.
Architecture, implementation, CI
PYTHON / DATA CONTRACTS
ExcelAlchemy (opens in a new tab)
A published library that turns Pydantic models into typed Excel import and export contracts.
Source, release, PyPI
AGENT WORKFLOWS
Codex Skills (opens in a new tab)
Reusable engineering procedures with evidence rules, validation, and clear side-effect boundaries.
Principles, validation contract
POINT OF VIEW
A good system leaves behind more than an answer.
Useful systems expose their contracts, make failures visible, and leave evidence that another engineer can inspect. That principle connects my backend work, data tooling, and agent workflows.