Open to opportunities

Steve Zhang

@stevezhang

Message

Senior software engineer building agentic AI and scalable payment infrastructure.

United States

Message

What I'm looking for

I’m looking for an opportunity to own end-to-end systems where agentic AI and high-volume reliability matter—building fast, auditable services, reducing cost/latency, and turning patterns into org-wide standards.

I’m a Senior Software Engineer with 12 years building agentic AI systems and high-throughput payment infrastructure across Google, Stripe, and Claros Analytics. At Google, I built and owned the Wallet fraud detection platform end to end, combining LangGraph multi-agent orchestration with real-time Vertex AI scoring pipelines and an analyst-facing React dashboard.

I lead full-stack, end-to-end delivery with a strong focus on reliability, auditability, and cost/latency wins—cutting analyst review time 40%, reducing inference cost 28%, and decomposing a 200k-line Java authorization monolith into Go gRPC microservices with Cloud Spanner to drive p99 latency from 150ms to 45ms. I also formalize patterns into org-wide standards, including an OpenTelemetry-based AI observability library adopted across Wallet AI services.

Experience

Work history, roles, and key accomplishments

Current

Senior Software Engineer

Current

Google

Feb 2019 - Present (7 years 4 months)

Built and owned the Google Wallet fraud detection platform, using LangGraph orchestration and Vertex AI scoring pipelines to cut analyst review time 40% with full audit compliance. Re-architected authorization and transaction processing to improve performance and reliability, including reducing p99 latency from 150ms to 45ms, absorbing 10x traffic growth, and cutting failover recovery to under 10

LangGraph Vertex AI fastAPI React TypeScript OpenTelemetry Cloud Spanner Kafka Streams

Software Engineer

Stripe

Aug 2016 - Jan 2019 (2 years 5 months)

Improved payment reliability by reducing double-charge incidents by 99.9%+ using Redis idempotency fingerprinting and event-driven AWS Lambda retry automation, cutting manual incident response 70%. Scaled billing/invoicing APIs to 99.95% uptime and built real-time React operations dashboards for payment failure visibility.

AWS Lambda Ruby on Rails Python React RSpec Pytest

Software Engineer

Claros Analytics

Jan 2014 - Jul 2016 (2 years 6 months)

Optimized a multi-tenant analytics platform by redesigning PostgreSQL schemas and query plans to improve dashboard load times 35% across 200+ enterprise tenants. Built core analytics APIs (Python/Django and Java/Spring Boot) with normalization, anomaly flagging, and scheduled report delivery, and reduced production regressions 40% by introducing a CircleCI test pipeline.

PostgreSQL Query Optimization Python Django Java Spring Boot CircleCi Pytest