Jianwei Li

Tech Stack: Go, go-zero, gRPC, Kafka, Redis, MySQL, MongoDB, Etcd, GORM, JWT, WebSocket, Docker, Kubernetes, Bitcoin JSON-RPC A Go-based digital asset trading platform supporting user accounts, wallets, deposits and withdrawals, spot trading, order placement, matching, K-line market data, order book updates, and scheduled backend jobs.

• Designed backend microservice boundaries across user center, trading center, market center, and job center modules using go-zero and gRPC.
• Built wallet capabilities including initialization, available/frozen balance management, deposit address maintenance, transaction records, and JWT-based user identity handling.
• Implemented spot order placement flows covering trading-pair validation, user trading status checks, limit/market order validation, balance checks, order persistence, and Kafka event publishing.
• Contributed to the matching engine, with isolated matching instances per trading pair and price-time priority for limit and market orders.
• Designed asynchronous consistency between orders and wallets: order creation triggers asset freezing, successful freezing enters the matching queue, and trade events update balances after execution.
• Built market data pipelines that persisted K-line data in MongoDB, cached latest prices in Redis, and pushed incremental updates through Kafka and WebSocket.
• Participated in BTC on-chain deposit scanning with Bitcoin Core JSON-RPC, including block parsing, transaction filtering, deposit detection, persistence, and event publishing.

Tech Stack: Java, Golang, Node.js, Spring Boot, Hibernate, Groovy, gRPC, internal RPC framework, Aurora MySQL, Redis, Pika, ElastiCache, AWS EC2/ECS/Lambda, Kafka, Kibana, Grafana A core DerbySoft hotel distribution connectivity project for direct integration between IHG's CRS and Booking.com. The system synchronized availability, rates, inventory, room types, rate plans, and promotions from IHG to Booking.com, while processing Booking.com reservations, cancellations, and modifications back to the hotel CRS.

• Designed and developed backend modules for the Booking.com and IHG CRS integration, covering ARI synchronization, room mapping, rate plan mapping, rate updates, promotion setup, and reservation delivery.
• Built channel protocol adaptation layers to transform DerbySoft/IHG standard data models into Booking.com-specific formats.
• Handled channel-specific differences in availability, rates, inventory, stay rules, cancellation policies, promotions, and reservation workflows.
• Implemented booking, cancellation, and modification processing to ensure reliable order delivery and consistent inventory state between Booking.com and the hotel CRS.
• Optimized high-concurrency service communication with gRPC and internal RPC frameworks to reduce latency across synchronization pipelines.
• Designed retry, idempotency, alerting, and traceable logging mechanisms to reduce the impact of third-party API instability, network timeouts, and duplicate messages.
• Supported product, implementation, QA, and overseas teams during Booking.com channel onboarding, including hotel activation, mapping, Genius Rate promotion setup, and end-to-end validation.

Tech Stack: Java, Golang, Spring Boot, Groovy, Kafka, Avro, JSON, Gradle, Jenkins, AWS ELB/EC2/RDS/DocumentDB, MongoDB, Redis, Elasticsearch, Kibana, Grafana Go Platform was DerbySoft's hotel connectivity middleware for suppliers and global distributors. It provided unified connectivity for hotel groups, independent hotels, PMS/CRS systems, OTAs, wholesalers, GDS, and TMC channels, covering availability, rates, inventory, reservations, protocol conversion, message orchestration, channel adaptation, data caching, logging, and configuration management.

• Developed the core connectivity framework for message consumption, protocol conversion, validation, and result write-back between hotel suppliers and distribution channels.
• Designed and implemented Kafka consumer modules that processed hotel data from upstream tasks and delegated Avro/JSON conversion to Groovy scripts.
• Built extensible channel adapters to convert CRS/PMS availability, rates, inventory, and room-type data into distributor-specific formats.
• Developed a Spring Boot starter that provided auto-configuration, service registration, unified authentication, log querying, and configuration management for connector services.
• Improved script execution performance through thread-pool tuning, compiled-script caching, batch processing, and fault isolation.
• Managed hotel, room-type, and channel mapping metadata with MongoDB/DocumentDB, while improving internal communication efficiency with Avro RPC and Kafka.
• Participated in CI/CD and release processes using Gradle and Jenkins, including packaging, automated builds, environment deployment, and version releases.

Tech Stack: Next.js 15, TypeScript, Prisma, MySQL, GitHub App, OpenRouter, Tailwind CSS, Docker A SaaS-style platform for translating GitHub repository documentation into multiple languages. Users can sign in with GitHub, import repositories, configure source and target languages, translate Markdown/MDX files with LLMs, and deliver results through pull requests.

• Designed and developed the MVP from scratch, including project structure, TypeScript types, UI styling, and local Docker setup.
• Modeled users, repositories, repository settings, translation tasks, translation files, API keys, and usage statistics with Prisma and MySQL.
• Integrated GitHub OAuth/GitHub App authorization and wrapped Octokit operations for repository import, file tree reading, branch creation, file writing, and pull request creation.
• Built the AI translation workflow for Markdown/MDX files, including file filtering, target path generation, OpenRouter calls, progress tracking, and failure recording.
• Implemented multi-model fallback and support for user-provided API keys to improve reliability and cost control.

Tech Stack: React 19, TypeScript, Vite, Tailwind CSS, Express 5, Prisma, SQLite, Socket.io, OpenRouter, Node-cron, Nodemailer An AI-powered hotspot monitoring platform that allows users to configure keywords, aggregate hot topics from multiple sources, and use LLMs for truthfulness checks, relevance analysis, importance scoring, and Chinese summary generation.

• Built the full-stack web platform with React and Express, including keyword management, hotspot lists, notifications, manual search, and real-time refresh.
• Implemented multi-source collection from Twitter/X, Bing, Hacker News, Sogou, Bilibili, Weibo, and other sources, with concurrent requests, deduplication, freshness filtering, and source-priority sorting.
• Integrated OpenRouter LLM capabilities for keyword expansion, content pre-matching, truthfulness evaluation, relevance scoring, importance judgment, and summary generation.
• Designed Prisma data models for keywords, hotspots, notifications, and system settings, using unique indexes to prevent duplicate hotspot records.
• Implemented scheduled checks every 30 minutes, manual triggering, WebSocket notifications, and email alerts for high-importance items.
• Packaged the monitoring and report-generation capability as an Agent Skill for use in Cursor, VSCode Copilot, Claude Code, and similar tools.

https://github.com/Leejaywell/github-i18n

https://github.com/Leejaywell/hot-track

https://github.com/Leejaywell/common-module

https://github.com/Leejaywell/ui-builder-backend https://github.com/Leejaywell/ui-builder-frontend

https://github.com/Leejaywell/blockchain-rs