The project is developed using the Golang, supports more than 40 typical sensitive data classifications, supports scanning of common relational databases and big data storage, and provides security recommendation reports for GDPR and data security laws. The algorithm is comprehensively implemented using operators such as column name keywords, regular expressions, and data dictionaries, and uses directed acyclic graphs to improve performance by 10 times. This product has been purchased by many security manufacturers, financial and overseas companies, and the design scheme has applied for a patent.

The project is implemented using Rust and Go. Go implements the control plane and is responsible for configuration conversion and distribution and operation and maintenance task scheduling. Rust implements the data plane and serves as a sidecar responsible for accessing business data traffic. The project implements MySQL data traffic processing, including read and write separation, sharding, access control, auditing and circuit breaker. The project currently has 270 stars, 10 active contributors, and 3 community enterprise users. The project address is github.com/database-mesh/pisanix

I have successively launched Helm, CloudFormation, Terraform and other IaC solutions for ShardingSphere, and designed an Operator that supports multiple CRDs and automatically deploys ShardingSphere and AWS RDS to form a distributed database. The Operator also supports chaos engineering based on ChaosMesh. In addition, it promotes the implementation of solutions such as community Wasm extensions, which improves the practicality of ShardingSphere in cloud environments and lowers the threshold for use. Project address: github.com/apache/shardingsphere-on-cloud

The project uses the Golang to implement the tenant portal and operation portal, public services and AWS-based cloud base, and designs shared or exclusive deployment solutions for different user groups. The bottom layer supports both EC2 and EKS, and allows users to open VPC Peering. Supports CloudFormation and AMI scenarios. In the first phase, the user system, artifact library system and release system have been implemented successively. Currently, the user system and artifact library services have been launched and have been running stably for 9 months, helping to improve delivery efficiency by 50%.

The project is implemented in Golang, supports Kubernetes native resource object management, supports cluster-level and instance-level Prometheus monitoring and chart drawing, supports ChartMesum application store docking with CICD, and has developed command line tools to support managed Kubernetes services in mainstream public clouds at home and abroad. This product has been recognized and selected by many domestic customers.

Led SRE team and responsible for building the company's AWS infrastructure from scratch, planning application transformation and cloud migration solutions, reducing transaction costs by 75%. Multi-cluster observability and alarm mechanisms are built based on Thanos, Prometheus, Grafana, CloudWatch, Lambda, etc. Implement a multi-lane elastic EKS cluster based on Spot instances, cooperate with the CICD test to automatically expand the cluster on demand, provide a multi-dimensional perspective on R&D performance, and be compatible with Helm's release management. In addition, I participated in the construction of engineering culture, including operation and maintenance work orders, emergency procedures, agile project management, etc., and completed technical upgrades.

As the main engineer, I built the company's Kubernetes container platform from 0 to 1, promoted business containerization transformation, and participated in the construction of Gitlab CICD pipeline to help improve delivery efficiency.

I mainly cooperate with business migration to the cloud, complete fine-grained resource overcommit control, support 618 and Double 11 promotions, build Ant OS base image pipelines, etc.