SOFTWARE ARCHITECTURE · FULL-STACK · AI
Java backend · ArchitectureFrontend / AI applications
I build Java business systems, from architecture to core implementation. My work spans orders, payments, telemarketing and messaging. I also develop frontend interfaces and contribute to AI applications.
01 / SELECTED WORK & ENGINEERING LAB
Each case covers the problem, my role and the design.
AIGC and quant work are personal offline prototypes, with no commercial delivery or live trading.
Responsible for architecture and core development for Yun Xiao Er and Yun Zhang Gui. Led the order module, developed the payment gateway, and used Netty connections and JPush for order notifications.
Used Kafka to handle bursts of delivery requests and a chain-of-responsibility design to apply rules across SMS, push notifications, IVR, and bot channels.
Led the move from a monolith to microservices. Addressed lock contention, timeouts and out-of-memory failures through JVM tuning, SQL indexes and Redis caching, and introduced job scheduling and central configuration.
Contributed to inquiry messaging and routing. The system receives inquiries through RabbitMQ, combines business context with retrieved knowledge to draft replies, and stores references and model-call details.
Contributed to a Dify workflow that lets operations staff query business data, statuses, and rules through a 360Teams bot, connecting intent recognition with internal APIs and replies.
An offline prototype built with AI assistance, using SQLite to record shot tasks and budgets and FFmpeg to inspect and process media. It explores failure recovery, duplicate execution and approval versioning.
A personal backtesting prototype built with AI assistance. LEAN calculates signals; a separate ledger models cash, fees, settlement and risk rules, with trade-by-trade checks. It has not been used for live trading.
The diagrams are simplified. Each case study details the design, implementation and my responsibilities.
02 / HOW I CAN HELP
Build a product, improve a system or solve a specific technical problem.
From requirements and design to implementation and handover.
Review service boundaries, APIs and data flows. Investigate slow queries, caching and asynchronous processing, then plan and implement the changes.
Build admin applications, order workflows and internal tools, including Vue interfaces, Java or Python services, databases and third-party integrations.
Connect models, knowledge retrieval and internal APIs for business queries. Prototype AIGC workflows to test task orchestration, recovery and human review.
Build backtests, account ledgers and reporting tools for research. Account for fees, settlement and risk rules so results can be checked trade by trade.
Project development / Phased delivery / Technical consulting
See the delivery process ↗LAB / INSIDE THE ARCHITECTURE
Click a step to read the design notes.
Explore business processing, content generation and backtesting.
A design walkthrough using order notifications and customer messaging to examine intake, rules, asynchronous work, and device acknowledgments.
Request acceptance, business completion, and device acknowledgment need separate records. Define what the response means and carry a business identifier into later tasks. Otherwise, a caller retrying after a timeout cannot easily tell whether the original request ran.
Simplified architecture walkthroughs drawn from project work and prototypes. Each case study explains its actual implementation scope.
03 / ENGINEERING TOOLKIT
Java is my main language, alongside frontend and AI work.
Personal research also uses Python, FFmpeg and LEAN.
Admin tools, UI development and API integration
Services, microservices and performance investigation
Databases, caching, search and message queues
Knowledge retrieval, business queries and routing
Builds, deployment, configuration and scheduling
Personal prototypes & research tools
04 / A LITTLE ABOUT ME
I started in Java backend development and have also worked on frontend interfaces. At Qifu Technology, I was responsible for telemarketing and messaging systems. I contributed to the AI customer service project in 2026 and also took part in operations self-service tools.
Outside work, I use AI assistance to build AIGC workflows and quant research tools, exploring task recovery, media checks and backtest validation. These are personal offline prototypes.
Read my résuméI've worked on monolith migrations and asynchronous messaging, including refund idempotency, transaction compensation, lock contention and timeouts.
Contributed to inquiry messaging and routing for AI customer service, working with business context, knowledge references and model-call records.
Led a team of around eight, covering task planning, code review and technical guidance, while coordinating releases with product, QA and operations.
05 / PROFESSIONAL EXPERIENCE
I started in Java backend development, later taking on architecture and team leadership. I also have frontend development experience.
Responsible for technical design, core development, and team delivery for telemarketing and messaging systems. Contributed to AI customer service and operations self-service projects. Led a team of around 8, with responsibilities including code review, technical guidance, hiring, and performance reviews.
Responsible for architecture and core development for Yun Xiao Er and Yun Zhang Gui. Led the order module, developed the payment gateway, and contributed to risk control.
Contributed to frontend development for the Youle Wang, You Sheng Huo, and You E Dai apps, improving initial load performance and list scrolling.
Responsible for Alipay external escrow transaction and batch refund modules. Designed refund idempotency checks and distributed locking, with compensation and retries for inconsistent transaction states.
Maintained and developed frontend features for AstraZeneca's portal.
06 / HOW WE WORK TOGETHER
Build in stages and review progress with working software.
Agree on source code, documentation and deployment handover at the start.
Understand the users, current system and priorities. Agree on scope and acceptance criteria.
Define technology choices, module responsibilities and APIs. Prototype any uncertain critical steps.
Get the main workflow running, add features and error handling, and demonstrate each stage.
Check features and failure scenarios, then hand over source code, documentation and deployment instructions.
Features, schedule, fees, and maintenance are agreed for each project after we discuss the requirements.
AVAILABLE FOR PROJECT COLLABORATION
Tell me what you're building and what needs solving.
We can also discuss your existing system, timeline and budget.
Email me directly or open a project inquiry on GitHub. GitHub inquiries are public; use email for detailed project materials.