Projects and engineering prototypes

CASE 01

Restaurant Orders & Payments

Baiweiyun (百味云) · Senior Software Engineer / Technical Team Lead · 2016.10 — 2018.04

An order can be processed while its notification is still waiting for a device acknowledgment.

Restaurant order and payment work at Baiweiyun, with RabbitMQ, Netty, JPush, application acknowledgments, and Redis controls for repeated notifications.

• Business architecture
• Orders / Payments
• RabbitMQ
• Netty
• Redis
• JPush

CASE 02

Enterprise Customer Messaging Platform

Qifu Technology (奇富科技) · Architecture & core development · 2020.05 — 2024.07

Kafka accepts the requests, a rule chain validates them, and channel strategies handle dispatch.

Architecture and core development of a Kafka messaging platform, with rule chains, channel dispatch, and carrier, weight, and quota-based SMS selection.

• Java
• Kafka
• Chain of responsibility
• Channel strategies
• Asynchronous processing
• Multiple providers

CASE 03

Outbound Calling Platform Modernization

Qifu Technology (奇富科技) · Development Lead · 2018.04 — 2026.03

The monolith needed service decomposition alongside fixes for lock contention, timeouts, and out-of-memory failures.

Led Qifu’s service decomposition and JVM, SQL, and cache work. The case also explains the system’s parallel queries and outbound validation.

• Java / Dubbo
• Microservices
• CompletableFuture
• JVM
• MySQL / Redis
• XXL-JOB / Apollo

CASE 04

AI Customer Support & Knowledge Retrieval

Qifu Technology (奇富科技) · Contributor to message processing and routing · 2026

I contributed to support-message processing and routing based on business context.

Contributed to Qifu’s AI support messaging and routing in 2026. The case explains ticket context, RabbitMQ task states, knowledge retrieval, and citations.

• Java 17
• Spring Boot 3
• LangChain4j
• RabbitMQ
• PostgreSQL
• RAG / Bailian

CASE 05

Dify Self-Service Assistant for Operations

Qifu Technology (奇富科技) · Project contributor · 2025.05 — 2026.03

Let operations query internal systems from 360Teams by translating questions into API calls.

A Dify project I contributed to: intent recognition, parameter extraction, and internal API queries for operations through a 360Teams bot.

• Dify
• Intent recognition
• Parameter extraction
• Business APIs
• Workflow orchestration
• 360Teams

CASE 06

AIGC Production Workflow

Personal R&D · AI-assisted · Offline engineering prototype

When a generation request times out, how can the task recover without reserving the same budget twice?

An AI-assisted personal offline prototype for recovery, budget accounting, and media review. Mock-provider validation only; no commercial delivery.

Offline workflow and mock-provider validation are complete. Real providers and generated-content quality have not been evaluated, and the prototype has not been used for commercial delivery.

• Python
• SQLite WAL
• FFmpeg / ffprobe
• State machines
• Budget reservations
• Task recovery

CASE 07

Quant Research & Backtesting

Personal research · AI-assisted · Simulated backtesting prototype

Reconcile signals, costs, and cash movements before interpreting a backtest.

An AI-assisted research prototype with LEAN signals and a Python ledger for costs and risk checks. Simulated backtests only, with no live trading.

Research and simulated backtesting only, with no brokerage connection or live orders. Backtest results do not establish future returns.

• C# / LEAN
• Python
• Independent capital ledger
• Cost modeling
• Risk rules
• Reproducible research