Yan Mao

Led end-to-end development of Deep Fleet (autonomous fleet coordination) and Project Eluna (agentic AI decision platform), enabling real-time optimization across global fulfillment centers via distributed systems, ML, and RAG-driven reasoning. Architected backend microservices with Java, Kafka, and gRPC to scale telemetry and decisions across billions of robot-hours. Engineered fault-tolerant services with Spring Boot and AWS Lambda to support mass robotic operations with high availability. Built ML models using PyTorch and Amazon SageMaker to improve congestion prediction and routing efficiency, and developed petabyte-scale data pipelines with Apache Spark and S3 for model training. Implemented observability with CloudWatch and X-Ray, and designed scalable deployment workflows using SageMaker Endpoints and Step Functions for low-latency inference. Created robust APIs with Java, Spring Boot, and API Gateway, enabling concurrent multi-facility operations. Led model evaluation, prompt en

Drove end-to-end validation and resilience for geo-federation storage solutions. Developed comprehensive test strategies for site outages, WAN partitions, and multi-VDc failover, and led validation of multi-site replication via Amazon S3 API and EMC Atmos, ensuring data consistency and recovery readiness. Implemented advanced data protection mechanisms including Reed-Solomon erasure coding (12+4, 10+2) and XOR encoding, and built automated frameworks (Python, Shell, Jenkins) to run daily regression suites. Designed automation workflows for replication policy validation and VDC endpoint synchronization, and collaborated with global support teams to troubleshoot complex multi-site issues for enterprise customers (Verizon, Swisscom). Managed test execution and defect tracking with TestRail and Jira, contributing to validation of next-gen ECS hardware (Gen 3 EX 300/EX 500) supporting up to 8 VDCs, enhancing federation capabilities and reliability.

Developed functional and regression test scenarios for geo-federation workflows, including site outages, network failures, and data recovery validation. Executed multi-interface testing using Amazon S3, OpenStack Swift, and EMC Atmos to ensure cross-storage compatibility and data integrity. Implemented automation scripts in Python and Shell to support ECS deployment, failure simulation, and regression testing; built end-to-end automation pipelines in Jenkins to maintain continuous validation of deployments and recovery workflows. Worked with engineering teams to reproduce and resolve customer issues, improving system reliability for enterprise clients.