ZhiGong Liu

Senior Software Engineer with 7+ years of experience, including 6 years at Amazon/AWS architecting fault-tolerant payment fee infrastructure (100+ TPS, <50ms p99 latency). I bridge high-scale distributed systems with AI, recently leading the design of a production LLM agent on AWS Bedrock. I specialize in exact-once processing, orchestrating reliable systems, and enforcing deterministic guardrails for AI integrations.

Software Development Engineer II - Amazon.com-Emerging Marketplace Payments - Seattle, WA/Vancouver, BC

(2021-08 - 2026-02)

• Applied AI & Orchestration: Architected a production-grade LLM advisory agent on Amazon Bedrock, driving self-service analytics by enabling stakeholders to simulate fee impacts via natural language; orchestrated parallel Lambda tool calls and enforced deterministic pre-compute guardrails to eliminate AI hallucinations.
• Performance & Scale: Architected the backend for the global Seller Pay Installment program, optimizing ECS microservices and DynamoDB to process 100+ TPS per region while maintaining strict <50ms p99 latency.
• Architecture & Business Impact: Generalized payment infrastructure into a centralized platform service, driving expansion across the US, Brazil, and Mexico to support 10M+ merchant adoptions.
• End-to-End Architecture: Led the design and implementation of a self-service merchant enrollment pipeline, reducing regional onboarding time by 50% and driving a 30% increase in Seller Adoption across emerging marketplaces.

Software Development Engineer - Amazon Web Services (AWS)-Control Tower - Seattle, WA, US

(2020-04 - 2021-07)

• Concurrency & Fault Tolerance: Optimized asynchronous infrastructure provisioning by implementing a Message Queue to serialize concurrent nested account requests, resolved critical race conditions and prevented unrecoverable CloudFormation failures, ensuring zero-touch onboarding for enterprise clients.
• Operational Excellence: Implemented event-driven orchestration pipeline with Lambda triggers to track account metadata and lifecycle events, paired with CloudWatch dashboards monitoring to reduce Mean Time to Resolution (MTTR) by 20%.

Full Stack Java Developer - Equifax LLC - St. Louis, MO, US

(2019-04 - 2020-04)

Migrated legacy financial data processing services to Google Kubernetes Engine (GKE) and architected automated Jenkins CI/CD pipelines, reducing manual deployment efforts by 20% while ensuring strict financial security compliance, aligned with SQL database schema migration and redesign.

Bachelor of Science - Computer Science - University of Wisconsin-Madison (2014 - 2018)