manutdlance

• - Responsible for full-stack development of overseas app backend business lines, including the operations backend. The frontend used Vue 3 + Element Plus + TypeScript, and the backend used the GoFrame framework. - Developed app server-side features, integrated Apple and Google payments, and implemented in-app purchases, coin top-ups, coin deduction/refunds, coin transaction records, user behavior monitoring, first-order discounts, template lists, categories, and related features. - Developed the operations backend as a full-stack system, including template configuration, package configuration, order monitoring, user management, template success-rate statistics, template synchronization, and configuration for video and image templates. - Integrated AIGC capabilities with official and third-party platforms, including Tencent Cloud VOD models such as PixVerse, Sora 2, Vidu, and Hailuo; Jimeng, Kling, Alibaba Cloud Bailian/Wanxiang, Happyhorse, Gemini, Veo 3.1, GPT, Grok, API Mart, and other API proxy platforms. Implemented features such as digital humans, pet daily-life content, and pet podcasts.

- Responsible for developing the financial reconciliation system, including system transaction reconciliation, merchant transaction reconciliation, asset snapshots, and related reports. - Responsible for developing and maintaining operations backend and merchant backend features, including meeting booking, asset reports, address book, webhook push, distributed API rate limiting, 2FA brute-force protection, IP blocking, account blocking, email anti-phishing code, sensitive field masking, large table migration, proactive account crediting, and internal system transfers. - Responsible for refactoring statistics-related modules, including daily trends and order volumes for deposits, withdrawals, currency exchange, and fund collection; daily transaction volume; merchant and coin rankings by asset and transaction volume; cost statistics by coin and chain; average transaction latency statistics, etc. - Integrated with third-party platforms, including Binance, OKX, and several fiat collection and payout platforms; also performed reverse engineering for checkout/cashier systems.

- Responsible for developing a distributed, high-concurrency IoV gateway with dynamic scalability, including both standard and epoll-based versions. With sufficient resources, the system can scale horizontally to support millions of concurrent connections. Current production concurrency is around 100K, with safe concurrency at approximately 200K. Work included socket programming, integration with T-Box devices, Prometheus, Redis, Kafka, Consul, Hystrix, daily maintenance, performance tuning, and architecture design for command uplink/downlink. - Responsible for developing the Connection Controller and Command Controller for data downlink, providing gRPC APIs for core long-connection information and remote command delivery. The service is a distributed, multi-node microservice component that supports horizontal scaling and high concurrency, with Hystrix used for circuit breaking and rate limiting. - Responsible for data parsing services that consume Kafka data, parse it, and provide gRPC APIs. The service works collaboratively across multiple nodes, consuming data from Kafka and storing processed data into HBase, Kafka, DataHub, and other systems. Supported protocols include GB17691, GB32960, 808, and customized JAC/JMC protocols. Daily data volume reaches hundreds of GB, making the service highly I/O-intensive and requiring strong write performance to avoid data latency under high concurrency. - Designed the Prometheus monitoring system, configured Consul for service discovery, and registered all components with Consul to build a monitoring system that supports horizontal node scaling. - Worked on Spark/PySpark big data development, including offline analysis tasks such as frequent stop location detection using the Geohash algorithm, daily mileage analysis, maximum trip radius analysis, national-standard alarms, battery health analysis, charging cycle count analysis, and anomaly clustering.

- Go + Python + Android (Java) development, crawling data from various applications. The backend was built with Golang, crawlers were developed in Python, and some tools were developed for Android. Main technologies included Selenium, Requests, JavaScript reverse engineering, Frida, Xposed, automated clicking, and packet capture. - Completed data crawling for Twitter, Weibo, Tencent Weiyun, Baidu Netdisk, Dropbox, OneDrive, email services (IMAP/POP/Exchange), Kuaishou and several short-video platforms, M3U8, Ele.me, Pinduoduo, QQ Sync Assistant, Skype, NetEase Mail, ZTO Express, WPS, Evernote, Zhihu, Lanzou Cloud, Tianyi Cloud, Meizu Cloud, Huawei Cloud, Xiaomi Cloud, Qunar, bank QR-code scanning scenarios, JD.com, and others. Experienced in handling Alibaba slider CAPTCHA, Geetest, NetEase Yidun, Tencent slider CAPTCHA, etc. Most implementations were based on packet capture protocol analysis and JavaScript reverse engineering, with collected data stored in MySQL. - Based on several technical blogs, implemented the conversion from the original emulator X86 instruction set to the ARM instruction set. See the related blog for details. This laid the foundation for the company’s emulator-based task system.